Navigation route processing method and device, storage medium and electronic equipment

Abstract

The application discloses a navigation route processing method and device, a storage medium and an electronic device, and can be applied to the field of maps. The method comprises the following steps: determining a group of navigation routes and a group of parking lots between a navigation starting point and a navigation ending point, obtaining a group of parking paths, each parking path comprising a combination of a navigation route and a parking lot; determining a vacant parking space feature for describing a historical vacant parking space of each parking lot in a target period, the target period being a period from the navigation starting point to the navigation ending point; determining a traffic flow state feature for describing a historical traffic flow state of each navigation route corresponding to the target period; inputting the vacant parking space feature of each parking lot and the traffic flow state feature of each navigation route into a target prediction model to obtain a target parking path, the target prediction model being used for predicting a selection probability of each parking path based on the correlation between the vacant parking space feature of each parking lot and the traffic flow state feature of each navigation route.

Description

Technical Field

[0001] This application relates to the field of computers, and more specifically, to a method and apparatus for processing navigation routes, a storage medium, and an electronic device. Background Technology

[0002] When a user needs vehicle navigation, they can input the starting and ending points on the navigation map, which will then plan a set of navigation routes for the user. The user can then choose from the set of routes or select a default route for vehicle navigation.

[0003] Currently, vehicle navigation processing can plan routes based on road traffic features and regional characteristics. However, features within the same area are very similar and do not change much over a long period of time, making it impossible to adapt to the rapidly changing road conditions in traffic scenarios. This results in unreasonable planned navigation routes, thus affecting users' travel efficiency.

[0004] It is evident that the navigation route processing methods in related technologies suffer from unreasonable navigation route planning due to the poor flexibility of path planning parameters. Summary of the Invention

[0005] This application provides a method and apparatus for processing navigation routes, a storage medium, and an electronic device, addressing the problem that navigation route processing methods in at least the related art suffer from unreasonable navigation route planning due to poor parameter flexibility in path planning.

[0006] According to one aspect of the embodiments of this application, a method for processing navigation routes is provided, comprising: determining a set of navigation routes between a navigation start point and a navigation end point, and determining a set of parking lots corresponding to the navigation end point to obtain a set of parking paths, wherein each parking path in the set of parking paths includes a combination of a navigation route in the set of navigation routes and a parking lot in the set of parking lots; determining the available parking space characteristics of each parking lot in the set of parking lots, wherein the available parking space characteristics of each parking lot are used to describe the historical available parking spaces of each parking lot during a target time period, the target time period being the time period from the navigation start point to the navigation end point; determining the traffic flow state characteristics of each navigation route in the set of navigation routes, wherein the traffic flow state characteristics of each navigation route are used to describe the historical traffic flow state of each navigation route corresponding to the target time period; inputting the available parking space characteristics of each parking lot and the traffic flow state characteristics of each navigation route into a target prediction model to obtain a target parking path output by the target prediction model, wherein the target prediction model is used to predict the selection probability of each parking path based on the correlation between the available parking space characteristics of each parking lot and the traffic flow state characteristics of each navigation route.

[0007] According to another aspect of the embodiments of this application, a navigation route processing apparatus is provided, comprising: a first determining unit, configured to determine a set of navigation routes between a navigation starting point and a navigation ending point, and to determine a set of parking lots corresponding to the navigation ending point, thereby obtaining a set of parking paths, wherein each parking path in the set of parking paths includes a combination of a navigation route in the set of navigation routes and a parking lot in the set of parking lots; and a second determining unit, configured to determine the available parking space characteristics of each parking lot in the set of parking lots, wherein the available parking space characteristics of each parking lot are used to describe the historical available parking spaces of each parking lot during a target time period, the target time period being from the navigation starting point... The system includes: a first unit for determining the time period from the point of travel to the navigation endpoint; a second unit for determining the traffic flow state characteristics of each navigation route in the set of navigation routes, wherein the traffic flow state characteristics of each navigation route are used to describe the historical traffic flow state of each navigation route corresponding to the target time period; and a third unit for inputting the available parking space characteristics of each parking lot and the traffic flow state characteristics of each navigation route into the target prediction model to obtain the target parking path output by the target prediction model, wherein the target prediction model is used to predict the selection probability of each parking path based on the correlation between the available parking space characteristics of each parking lot and the traffic flow state characteristics of each navigation route.

[0008] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, wherein a computer program is stored in the computer program, and the computer program is configured to execute the above-described navigation route processing method when it is run.

[0009] According to another aspect of the embodiments of this application, a computer program product or computer program is provided, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the navigation route processing method described above.

[0010] According to another aspect of the embodiments of this application, an electronic device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the above-described navigation route processing method through the computer program.

[0011] In this embodiment, the navigation route planning method based on road traffic flow and available parking spaces at the navigation endpoint involves first determining a set of navigation routes between the navigation start point and the navigation endpoint, and a set of parking lots corresponding to the navigation endpoint, resulting in a set of parking paths. Each parking path is a combination of a navigation route and a parking lot. Then, for each navigation route, the traffic flow state characteristics of each route are determined, i.e., historical traffic flow changes corresponding to the navigation time period (i.e., the target time period), and the available parking space characteristics of each parking lot are determined, i.e., historical available parking spaces of each parking lot during the navigation time period. Based on the traffic flow state characteristics of each navigation route and the available parking space characteristics of each parking lot, the method is further refined. This method uses a pre-trained target prediction model to predict the probability of selecting each parking route, thereby selecting the target parking route from a set of parking routes. Since the traffic flow status and parking lot availability change over time, it is highly flexible and can analyze and discover change patterns in real time, leading to better route planning. At the same time, by incorporating the availability of parking spaces, it can better predict suitable parking lots when arriving at the navigation destination, facilitating the search for available parking spaces and shortening the time spent searching for parking spaces. This achieves the technical effect of improving the rationality of navigation route planning, thus solving the problem of unreasonable navigation route planning caused by the poor flexibility of path planning parameters in related technologies. Attached Figure Description

[0012] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0013] Figure 1 This is a schematic diagram of an application environment for an optional navigation route processing method according to an embodiment of this application;

[0014] Figure 2 This is a flowchart illustrating an optional navigation route processing method according to an embodiment of this application;

[0015] Figure 3 This is a schematic diagram of an optional navigation route processing method according to an embodiment of this application;

[0016] Figure 4 This is a schematic diagram of an optional navigation route processing method according to an embodiment of this application;

[0017] Figure 5 This is a schematic diagram of an optional encoder model according to an embodiment of this application;

[0018] Figure 6 This is a partial schematic diagram of an optional prediction model according to an embodiment of this application;

[0019] Figure 7 This is a partial schematic diagram of another optional prediction model according to an embodiment of this application;

[0020] Figure 8 This is a partial schematic diagram of another optional prediction model according to an embodiment of this application;

[0021] Figure 9 This is a flowchart illustrating another optional navigation route processing method according to an embodiment of this application;

[0022] Figure 10 This is a structural block diagram of an optional navigation route processing device according to an embodiment of this application;

[0023] Figure 11 This is a schematic diagram of the structure of an optional electronic device according to an embodiment of this application;

[0024] Figure 12 This is a structural block diagram of a computer system for an optional electronic device according to an embodiment of this application. Detailed Implementation

[0025] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0026] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0027] According to one aspect of the present invention, a method for processing navigation routes is provided. As an optional implementation, the above-described method for processing navigation routes can be applied, but is not limited to, to applications such as... Figure 1 The application scenarios shown are as follows. In, for example... Figure 1 In the application scenario shown, the process of navigating a vehicle through a map application on terminal device 102 is as follows: the map application obtains the navigation start point, navigation destination, and navigation time of the vehicle navigation, and transmits them to server 106 via the network. Server 106 plans the navigation route based on the navigation start point, navigation destination, and navigation time, determines the navigation route and parking lots near the navigation destination, and sends the navigation information to terminal device 102 for display.

[0028] The process of server 106 planning navigation routes can be as follows:

[0029] Step S102: Receive a navigation request, which includes the navigation start point, navigation destination, and navigation time. Step S104: Perform route planning based on the navigation start point, navigation destination, and navigation time to obtain navigation information including the navigation route and parking lots near the navigation destination. Step S106: Transmit the navigation information to the terminal device to display the navigation route and parking lots near the navigation destination. The above is merely an example, and this embodiment does not impose any limitations.

[0030] Optionally, the terminal device 102 may include, but is not limited to, at least one of the following: mobile phones (such as Android phones, iOS phones, etc.), laptops, tablets, handheld computers, mobile internet devices (MIDs), tablets, desktop computers, smart home appliances, in-vehicle devices, etc. The network 110 may include, but is not limited to, wired networks and wireless networks. The wired network includes local area networks (LANs), metropolitan area networks (MANs), and wide area networks (WANs). The wireless network includes Bluetooth, Wireless Fidelity (WIFI), and other networks that enable wireless communication. The server 106 may be a single server, a server cluster consisting of multiple servers, or a cloud server. The above is merely an example, and no limitations are imposed in this embodiment.

[0031] As an optional example, this embodiment does not limit the execution subject of the above steps S102 to S106. For example, the above steps S102 to S106 can all be executed on the terminal device 102. This embodiment does not limit this.

[0032] As an optional implementation, taking the navigation route processing method in this embodiment executed by server 106 as an example, Figure 2 This is a flowchart illustrating an optional navigation route processing method according to an embodiment of this application, as shown below. Figure 2 As shown, the process of processing this navigation route may include the following steps:

[0033] Step S202: Determine a set of navigation routes between the navigation start point and the navigation end point, and determine a set of parking lots corresponding to the navigation end point to obtain a set of parking paths. Each parking path in the set of parking paths includes a combination of a navigation route in the set of navigation routes and a parking lot in the set of parking lots.

[0034] The navigation route processing method in this embodiment can be applied to fields such as urban road traffic and urban resource allocation, including traffic route planning, map route navigation, and urban resource management and allocation. For example, in urban resource allocation, by mining and analyzing the state distribution of traffic areas, more traffic resources and parking spaces can be allocated and rationally distributed in areas with high traffic volume, reducing traffic pressure and parking difficulties in these areas. In map route navigation applications, by identifying and predicting traffic conditions between the origin and destination in real time, and using road traffic data, more effective route references can be provided based on traffic congestion. By predicting traffic congestion and optimizing routes, the accuracy of navigation routes can be improved. At the same time, the fastest and most convenient parking location can be selected, reducing travel time and improving efficiency. In addition, all or part of the navigation route processing method in this embodiment can be applied to applications related to traffic flow area state prediction and parking route planning.

[0035] Taking map route navigation as an example, to improve the rationality of route planning, it is possible to combine regional traffic conditions and parking information near the navigation destination for route planning. For example, route planning can be performed using at least one of the following methods: prediction models based on deep learning networks, prediction models based on clustering and genetic algorithms, and prediction models based on nearest-neighbor traffic congestion indices.

[0036] For prediction models based on deep learning networks, a new feature set can be generated by learning hidden layer parameters that can represent deep features of the data from unlabeled datasets using the autoencoder network method of deep learning. Softmax (i.e., normalized exponential function) regression is then applied to learn the new labeled feature set to generate a prediction classifier. The prediction model can make polymorphic predictions on parking spaces and traffic congestion.

[0037] However, supervised classification models that predict whether a region is congested require labeling a portion of the samples. Furthermore, since road traffic features are highly correlated with regional features, the features are very similar within the same region and do not change much over a long period of time, making it difficult to adapt to the rapidly changing road conditions in traffic route scenarios.

[0038] For prediction models based on clustering and genetic algorithms, the historical traffic congestion and parking space datasets are first clustered using the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) clustering algorithm. Then, using the collected road traffic congestion and parking space information, the temporal association rules of roads are mined according to the genetic algorithm (a type of deep learning algorithm), thereby predicting the traffic congestion status and parking spaces at future times.

[0039] However, the key problem that deep learning solves lies in how to construct the network relationship between features related to parking scenarios. Existing methods often do not integrate parking space information and traffic volume features, nor do they consider mining the sequence correlation relationship between features.

[0040] For the nearest neighbor-based traffic congestion index prediction model, by constructing the traffic congestion index details as the prediction objective function, a K-nearest neighbor-based urban road traffic congestion index prediction model is established, and the model's state vector, distance calculation method, and prediction value calculation method are determined, thereby enabling supervised prediction.

[0041] However, the nearest neighbor-based traffic congestion index prediction model is a prediction model built on statistical learning and statistical indicators. Theoretically, kernel technology can map the original feature space to a new feature space, but the choice of which kernel to use for mapping needs to be set manually. Furthermore, statistical learning and artificial neural networks have a single form of expression, which is "attribute-value" and cannot meet the requirements of representing the relationship between multidimensional features. In addition, in order to ensure that the problem to be solved meets the properties of statistics or to simplify the problem, statistical learning techniques will make some assumptions that do not conform to the real situation. These assumptions are difficult to hold under actual conditions, thus affecting the accuracy of the prediction.

[0042] To at least partially solve the above-mentioned technical problems, in this embodiment, route planning is performed by combining historical traffic flow data and historical changes in available parking spaces near the navigation destination. The historical traffic flow data and historical changes in available parking spaces correspond to the navigation time, and both are related to the navigation route planning. Referring to both for navigation route planning can improve the rationality of the planned navigation route.

[0043] When navigation route planning is required, the user (identifiable as the target object by the terminal device) can input or select the navigation start point, navigation destination, and navigation time (which can be the departure time triggered from the navigation start point or the arrival time to the navigation destination) through the terminal device's map application. This is then triggered by activating a navigation control (e.g., a button). For the terminal device, the destination device can obtain the navigation start point, navigation destination, and navigation time set in the map application and, in response to the trigger operation performed on the navigation control, send a navigation request carrying the navigation start point, navigation destination, and navigation time to the server. Alternatively, for offline maps, the terminal device can directly generate a navigation trigger command to initiate map navigation from the navigation start point to the navigation destination that meets the navigation time requirements.

[0044] Taking the navigation route planning performed by the server as an example, the server can receive navigation requests, extract the navigation start point, navigation end point and navigation time from them, and plan the navigation route according to the navigation start point and navigation end point, or the navigation start point, navigation end point and navigation time, and determine a set of navigation routes between the navigation start point and the navigation end point. Each navigation route can be a route from the navigation start point to the navigation end point.

[0045] The server can also determine a group of parking lots corresponding to the navigation endpoint. Each parking lot in this group is a parking lot that vehicles can enter after arriving at the navigation endpoint, that is, enter the parking lot for parking. It can be any parking lot or a parking lot that meets certain conditions with respect to the navigation endpoint, such as time conditions, distance conditions, etc. In this embodiment, the method of determining a group of parking lots is not limited.

[0046] For example, such as Figure 3 As shown, multiple navigation routes (e.g., route 1-route 3) are determined between the navigation start point and the navigation end point, as well as multiple parking lots (e.g., parking lot AF) near the navigation end point.

[0047] A set of parking routes can be obtained based on a defined set of navigation routes and a set of parking lots. Each parking route in the set of parking routes is a combination of a navigation route from the set of navigation routes and a parking lot from the set of parking lots.

[0048] For example, given the aforementioned navigation routes 1 to 3 and parking lots A to F, 18 parking routes can be obtained. From these, a final parking route can be selected, namely, the combination of route 2 and parking lot E, such as... Figure 4 As shown.

[0049] Step S204: Determine the available parking space characteristics of each parking lot in a set of parking lots. The available parking space characteristics of each parking lot are used to describe the historical available parking spaces of each parking lot during the target time period, which is the time period from the navigation starting point to the navigation destination.

[0050] For each parking lot in a set of parking lots, the characteristics of available parking spaces for each parking lot can be determined. These characteristics describe the historical availability of parking spaces for each parking lot during a target time period. They can be the number of historically available parking spaces for each parking lot during the target time period, the change in historically available parking spaces for each parking lot during the target time period, or other characteristics. Optionally, the characteristics of available parking spaces for each parking lot can be the change in available parking spaces for each parking lot during the target time period, or the difference between the number of available parking spaces at the end of the target time period and the number of available parking spaces at the beginning of the target time period.

[0051] Optionally, the target time period can be one of a set of time periods obtained by dividing a time period according to time intervals. Here, a time period can be a day or other time periods. The time interval can be a preset time interval, such as one hour, five minutes, or other time intervals. The time interval can also be dynamically set. For example, the time interval for dividing a time period can be determined based on the time taken for each navigation route to travel from the navigation start point to the navigation end point. For example, if the time required for each navigation route to travel from the navigation start point to the navigation end point is half an hour, the time interval can be set to one hour. Or, if the time taken to travel from the navigation start point to the navigation end point spans two hours, the time interval can be set to two hours, and so on.

[0052] Since the availability of parking spaces in each parking lot changes over different time periods, determining the availability characteristics of each parking lot includes: determining the availability characteristics of each parking lot in each time period across multiple time periods. That is, the availability characteristics of each parking lot in each time period are used to describe the historical availability of parking spaces in each parking lot during the target time period of each time period.

[0053] Optionally, parking space availability data at various time intervals (e.g., every minute) in a parking lot can be identified and obtained through terminal devices such as traffic cameras, and the data can be standardized. For example, a correspondence between parking lot, time period, and available parking space can be set, or other standardization methods can be used. This embodiment does not limit this.

[0054] Step S206: Determine the traffic flow state characteristics of each navigation route in a set of navigation routes, wherein the traffic flow state characteristics of each navigation route are used to describe the historical traffic flow state of each navigation route corresponding to the target time period.

[0055] For each navigation route in a set of navigation routes, based on navigation time, such as departure time and end time, the time to travel from the navigation start point to the navigation end point along each navigation route can be determined. This can be the time to move to a preset node (e.g., an intersection) on each navigation route, thereby determining the time period for traveling from the navigation start point to the navigation end point along each navigation route, i.e., the target time period. Based on the target time period, the traffic flow state characteristics of each navigation route can be determined. The traffic flow state characteristics of each navigation route are used to describe the traffic flow state of each navigation route corresponding to the target time period. Here, the traffic flow state can be the historical traffic flow state, which can be the traffic flow state within one time period (e.g., the traffic flow state within the previous time period of the current time period) or the traffic flow state within multiple time periods (e.g., multiple time periods before the current time period).

[0056] Optionally, the traffic flow status characteristics of each navigation route may include, but are not limited to, at least one of the following: traffic volume corresponding to the target time period for each navigation route, and the traffic flow change for each navigation route corresponding to the target time period. Furthermore, the traffic flow status characteristics of each navigation route may also include other characteristics that can affect traffic flow, such as whether it is rush hour or a holiday. Since the traffic flow status of each navigation route changes over different time periods, determining the traffic flow status characteristics of each navigation route includes: determining the traffic flow status characteristics of each navigation route in each of multiple time periods; that is, the traffic flow status characteristics of each navigation route in each time period are used to describe the traffic flow status of each navigation route corresponding to the target time period in each time period.

[0057] It should be noted that parking space availability data at various time intervals (e.g., every minute) and 24-hour traffic flow data for various road segments in the city can be obtained through terminal devices such as traffic cameras. This data can then be standardized, and steps S204 to S206 can be executed based on the results of the standardization process. For example, the correspondence between navigation routes and parking lots can be shown in Table 1.

[0058] Table 1

[0059] Route 1 Parking Lot A Route 1 Parking Lot B Route 1 Parking lot C Route 1 Parking Lot D Route 1 Parking lot E Route 1 Parking lot F Route 2 Parking Lot A … … Route n Parking lot N

[0060] There is a correspondence between different routes and parking lots. The availability of 24-hour parking spaces for Route 1 is shown in Table 2.

[0061] Table 2

[0062] Route 1 Parking Lot A 0 o'clock 20 Route 1 Parking Lot A 1 o'clock 15 Route 1 Parking Lot A 2 o'clock 26 Route 1 Parking Lot A 3 o'clock 9 Route 1 Parking Lot A 4 o'clock 5 Route 1 Parking Lot A 5 o'clock 10 Route 1 Parking Lot A 6 o'clock 50 Route 1 Parking Lot A 7 o'clock 76 Route 1 Parking Lot A 8 o'clock 90 Route 1 Parking Lot A 9 o'clock 120 Route 1 Parking Lot A 10:00 100 Route 1 Parking Lot A 11:00 70 Route 1 Parking Lot A 12 o'clock 90 Route 1 Parking Lot A 13:00 80 Route 1 Parking Lot A 2 PM 75 Route 1 Parking Lot A 3 PM 130 Route 1 Parking Lot A 16:00 60

[0063] Table 2 (continued)

[0064] Route 1 Parking Lot A 5 PM 30 Route 1 Parking Lot A 6 PM 154 Route 1 Parking Lot A 7 PM 170 Route 1 Parking Lot A 8 PM 164 Route 1 Parking Lot A 21:00 90 Route 1 Parking Lot A 22:00 70 Route 1 Parking Lot A 11 PM 40

[0065] For the same parking lot, the number of available parking spaces changes dynamically at different times. Here, available parking spaces refer to those available at a specific hour (or other time). The traffic flow mapping relationship between Route 1 and each parking lot is shown in Table 3.

[0066] Table 3

[0067]

[0068] Table 3 (continued)

[0069]

[0070] Based on the mapping table of 24-hour routes and the number of available parking spaces in each parking lot, the changes in the number of available parking spaces in each parking lot within a time interval can be determined, as shown in Table 4.

[0071] Table 4

[0072]

[0073] Table 4 (continued)

[0074]

[0075] For example, you can obtain the set time interval, such as setting it to every minute / hour, and calculate the difference between the traffic flow in the next minute / hour and the available parking spaces and traffic flow in the previous minute / hour to obtain the change value within that time interval.

[0076] Step S208: Input the available parking space features of each parking lot and the traffic flow status features of each navigation route into the target prediction model to obtain the target parking path output by the target prediction model. The target prediction model is used to predict the selection probability of each parking path based on the correlation between the available parking space features of each parking lot and the traffic flow status features of each navigation route.

[0077] Based on the traffic flow characteristics of each navigation route and the availability of parking spaces in each parking lot, a target parking route can be selected from a set of parking routes. There are several ways to select a target parking route from a set of parking routes. For example, it can be based on the traffic flow characteristics of each navigation route to predict congestion during movement along each route, determining the likelihood of vehicle congestion and thus obtaining the probability of congestion along each navigation route; or it can be based on the availability of parking spaces in each parking lot to predict whether there are available spaces upon arrival, determining the likelihood of finding an available space and thus obtaining the probability of available spaces in each parking lot; the combination of the navigation route with the lowest probability of congestion and the parking lot with the highest probability of available spaces is determined as the target parking route.

[0078] Optionally, traffic flow along the navigation route can affect the availability of parking spaces in the parking lot (for example, if the traffic volume along the current navigation route is high, the availability of parking spaces in the parking lot is likely to decrease), and the availability of parking spaces in the parking lot can also affect traffic flow along the navigation route (for example, fewer parking spaces in the parking lot can lead to longer times for vehicles to find parking spaces, increasing the likelihood of traffic congestion). To address this, a target prediction model can be pre-built. This model is a network model built with parking path classification probabilities as its output. The final output is the parking path classification probability, which, based on the correlation between the availability of parking spaces in each parking lot and the traffic flow characteristics of each navigation route, predicts the probability of selecting each parking path—that is, the probability of selecting each combination of navigation route and parking lot. The parking path with the highest selection probability is then output, thus obtaining the target parking path output by the target prediction model.

[0079] The embodiments provided in this application determine a set of navigation routes between a navigation start point and a navigation end point, and determine a set of parking lots corresponding to the navigation end point, thus obtaining a set of parking paths. Each parking path in the set of parking paths includes a combination of one navigation route from the set of navigation routes and one parking lot from the set of parking lots. The available parking space characteristics of each parking lot in the set of parking lots are determined, wherein the available parking space characteristics of each parking lot describe the historical available parking spaces of each parking lot during a target time period, where the target time period is the time from the navigation start point to the navigation end point. The traffic flow state characteristics of each navigation route in the set of navigation routes are also determined. The traffic flow state features of each navigation route are used to describe the historical traffic flow state of each navigation route corresponding to the target time period. The available parking space features of each parking lot and the traffic flow state features of each navigation route are input into the target prediction model to obtain the target parking path output by the target prediction model. The target prediction model is used to predict the selection probability of each parking path based on the correlation between the available parking space features of each parking lot and the traffic flow state features of each navigation route. This solves the problem of unreasonable navigation route planning caused by the poor parameter flexibility of path planning in the navigation route processing methods of related technologies, and improves the rationality of navigation route planning.

[0080] As an optional approach, the available parking space characteristics of each parking lot and the traffic flow status characteristics of each navigation route are input into the target prediction model to obtain the target parking path output by the target prediction model, including:

[0081] S11, input the available parking space features of each parking lot and the traffic flow status features of each navigation route into the feature fusion layer of the target prediction model to obtain a set of first fusion features output by the feature fusion layer;

[0082] S12, input the available parking space features of each parking lot, the traffic flow status features of each navigation route and a set of first fusion features into the target neural network layer of the target prediction model to obtain the selection probability of each parking path output by the target neural network layer;

[0083] S13, determine the parking path with the highest selection probability from a set of parking paths as the target parking path.

[0084] When selecting a parking route from a set of parking routes, in order to explore the correlation and causal relationship between traffic flow status and parking space changes, a feature fusion layer of the target prediction model can be used to fuse the traffic flow status features of each navigation route and the available parking space features of each parking lot to obtain a set of first fused features. The feature fusion layer can be a pre-trained feature extraction network, such as CNN (Convolutional Neural Network), RNN (Recurrent Neural Network), or other neural network layers used for feature fusion. This embodiment does not limit this.

[0085] Optionally, a target parking route can be selected from a set of parking routes based on the traffic flow status characteristics of each navigation route, the available parking space characteristics of each parking lot, and a set of first fusion features. For example, the available parking space characteristics of each parking lot, the traffic flow status characteristics of each navigation route, and a set of first fusion features can be input into the target neural network layer of the target prediction model, thereby combining the traffic flow status characteristics of each navigation route, the available parking space characteristics of each parking lot, and a set of first fusion features to determine the selection probability of each parking route.

[0086] After obtaining the selection probability of each parking path, the target prediction model can output the parking path with the highest selection probability among a set of parking paths. The parking path with the highest selection probability is the target parking path.

[0087] Through the embodiments provided in this application, by fusing the traffic flow status characteristics of each navigation route and the available parking space characteristics of each parking lot, the correlation and causal relationship between traffic flow status and parking space changes can be explored, thereby improving the rationality of parking route selection.

[0088] As an alternative approach, the available parking space features of each parking lot and the traffic flow status features of each navigation route are input into the feature fusion layer of the target prediction model to obtain a set of first fusion features, including:

[0089] S21, input the available parking space features of each parking lot and the traffic flow status features of each navigation route into the encoder model of the target prediction model to obtain a set of first fusion features output by the encoder model.

[0090] In this embodiment, the feature fusion layer used for feature fusion can be an encoder model, i.e., a Transformer model. The traffic flow state features of each navigation route and the available parking spaces of each parking lot can be input into the encoder model to obtain a set of first fused features output by the encoder model.

[0091] Here, compared to CNN, Transformer can acquire global information. At the same time, Transformer improves upon the slow training of RNN by utilizing self-attention to achieve fast parallelism. The structure of Transformer is as follows: Figure 5 As shown.

[0092] like Figure 5 As shown, the Transformer model has a Multi-Head Self Attention structure, which contains multiple layers of attention with identical structures but different weight matrices. This structure prevents the model from focusing on only a portion of the model's features. Through the multi-head design, each head focuses on different features, so the model as a whole focuses on more features. It fuses the constructed traffic flow state features (e.g., the traffic congestion data features mentioned below), learns the correlation between the traffic flow state features of each navigation route (e.g., traffic flow change sequence) and the available parking space features of each parking lot (e.g., parking space change sequence), allowing the model to learn multiple different information from different perspectives and then fuse them.

[0093] The embodiments provided in this application demonstrate that feature fusion using an encoder model can improve the comprehensiveness of information acquisition, accelerate model training, and enhance the comprehensiveness of feature fusion.

[0094] As an optional approach, the available parking space features of each parking lot, the traffic flow status features of each navigation route, and a set of first fusion features are input into the target neural network layer of the target prediction model to obtain the selection probability of each parking path output by the target neural network layer, including:

[0095] S31. Generate a set of parking sequence features for each parking lot based on the characteristics of available parking spaces.

[0096] The available parking space characteristics of each parking lot are used to describe the historical available parking spaces of each parking lot during the target time period. These characteristics can be the historical available parking spaces at different times within the target time period, or the difference between the available parking spaces at the end and beginning of the target time period. In this embodiment, the available parking space characteristics used for parking route selection can be the available parking space characteristics of each parking lot over multiple time periods (historical time periods, such as the past two days, the past three days, etc.). Based on the available parking space characteristics of each parking lot over multiple time periods, a set of target parking lots within a group of parking lots can be determined, along with the trend of available parking space changes for each target parking lot, thus obtaining the parking sequence characteristics of a group of parking lots.

[0097] Here, the target parking lot refers to a group of parking lots in which the number of parking lots exhibiting the same trend of changing available parking spaces within a target time period of multiple time periods is greater than a first preset number. The trend of changing available parking spaces can be varied, including but not limited to at least one of the following: an increasing trend in available parking spaces, a decreasing trend in available parking spaces, or a stable trend in available parking spaces (no change in the number of parking spaces). The first preset number can be a preset value, determined by the product of the number of time periods encompassed by the multiple time periods and a first minimum support. For example, if the number of time periods encompassed by the multiple time periods is 2 and the first minimum support is 0.6, then the first preset number is 1.2.

[0098] For example, based on data obtained from different time intervals within each time period of parking spaces near the destination, the changes in available parking spaces for each time interval can be obtained, as shown in Table 4. Then, the increasing or decreasing trends (i.e., the positive or negative difference in traffic flow within the time interval) are marked with symbols: an increase (positive difference) is marked as 1, a decrease (negative difference) is marked as -1, and no change (0 difference) is marked as 0. The changes in the number of available parking spaces in each parking lot within each time interval after processing are shown in Table 5.

[0099] Table 5

[0100]

[0101] Based on Table 5, the trends of increase and decrease in available parking spaces at each parking lot along the route, divided by time intervals, can be determined, as shown in Table 6:

[0102] Table 6

[0103]

[0104] Furthermore, based on the changing trends of each parking lot along the route within time intervals, a sequence of changes in available parking spaces (or parking space change sequences) can be constructed for each parking lot. For example, for each parking lot along Route 1, the sequence of changes in available parking spaces constructed according to Table 6 is shown in Table 7.

[0105] Table 7

[0106]

[0107] Based on the increase or decrease trend of available parking spaces in each parking lot during the same period over multiple time periods, a set of target parking lots that meet the conditions can be selected from a set of parking lots.

[0108] S32, based on the traffic flow state characteristics of each navigation route, generate the traffic sequence characteristics of each navigation route.

[0109] Each navigation route can include multiple road segments. Different navigation routes may contain the same number of road segments or different numbers of road segments. For example, the relationship between routes and road segments can be shown in Table 8.

[0110] Table 8

[0111] Route 1 Section A Route 1 Section B Route 1 Section C Route 1 Section D Route 1 Section E Route 1 Section F Route 2 Section A … … Route n Section N

[0112] For a navigation route, such as the current navigation route, traffic sequence features can be generated based on the traffic flow status characteristics of the current navigation route. These traffic flow status features can be traffic flow within a time period, or the difference between the traffic flow in one time period and the traffic flow in another time period (e.g., the previous time period). The traffic flow data can be shown in Table 9.

[0113] Table 9

[0114] Route 1 Section A 0 o'clock 20 Route 1 Section A 1 o'clock 15 Route 1 Section A 2 o'clock 26 Route 1 Section A 3 o'clock 9 Route 1 Section A 4 o'clock 5 Route 1 Section A 5 o'clock 10 Route 1 Section A 6 o'clock 50

[0115] Table 9 (continued)

[0116] Route 1 Section A 7 o'clock 76 Route 1 Section A 8 o'clock 90 Route 1 Section A 9 o'clock 120 Route 1 Section A 10:00 100 Route 1 Section A 11:00 70 Route 1 Section A 12 o'clock 90 Route 1 Section A 13:00 80 Route 1 Section A 2 PM 75 Route 1 Section A 3 PM 130 Route 1 Section A 16:00 60 Route 1 Section A 5 PM 30 Route 1 Section A 6 PM 154 Route 1 Section A 7 PM 170 Route 1 Section A 8 PM 164 Route 1 Section A 21:00 90 Route 1 Section A 22:00 70 Route 1 Section A 11 PM 40

[0117] In this embodiment, the traffic flow state characteristics used for parking route selection can be the traffic flow state characteristics of each navigation route over multiple time periods. Based on the traffic flow state characteristics of each navigation route over multiple time periods, a set of target road segments and the traffic flow change trend of each target road segment can be determined for each navigation route, thereby obtaining the traffic sequence characteristics of each navigation route.

[0118] Here, the target road segment refers to a road segment within a navigation route where, within a target time period across multiple time periods, the number of road segments exhibiting the same traffic flow trend exceeds a second preset number. The traffic flow trend can be varied, including but not limited to at least one of the following: an increasing trend in traffic flow, a decreasing trend in traffic flow, or a stable trend in traffic flow (no change in traffic flow). The second preset number can be a preset value, determined by multiplying the number of time periods encompassed across the multiple time periods by a second minimum support. For example, if the number of time periods encompassed across the multiple time periods is 2, and the first minimum support is 0.6, then the second preset number is 1.2.

[0119] In this embodiment, the method of selecting a set of target road segments from multiple road segments of each navigation route and determining the traffic flow change trend of each target road segment is similar to the aforementioned method of selecting a set of target parking lots from a set of parking lots and determining the traffic flow change trend of each road segment, and will not be described in detail here.

[0120] S33, the parking sequence features, the traffic sequence features of each navigation route and a set of first fusion features are input into the feature extraction layer in the target neural network layer for feature extraction, and a set of second fusion features is obtained from the output of the feature extraction layer.

[0121] The target neural network layer may contain a feature extraction layer, which can perform deep feature extraction on the input parking sequence features, traffic sequence features of each navigation route and a set of first fusion features to obtain a set of fusion features, namely, a set of second fusion features.

[0122] S34, a set of second fusion features is input into the output layer of the target neural network layer to obtain the output result of the output layer, wherein the output result of the output layer is used to represent the selection probability of each parking path.

[0123] The obtained set of second fusion features can be input into the output layer of the target neural network to obtain the output result of the output layer, which represents the selection probability of each parking path. For example, the sigmoid function can be used as the output layer, and the loss function can be the cross-entropy loss function, as shown in formula (1):

[0124]

[0125] Where L is the loss function, x is the model input, y is the model output, and N is the number of outputs.

[0126] By using the embodiments provided in this application, by determining the parking sequence characteristics and the traffic sequence characteristics of each navigation route, and by using the determined parking sequence characteristics and the traffic sequence characteristics of each navigation route to select a parking route, the influence of noise characteristics can be reduced and the rationality of parking route selection can be improved.

[0127] As an optional approach, the available parking space characteristics of each parking lot include a sequence of available parking space changes corresponding to each of multiple time periods. Each element in the sequence of available parking space changes corresponding to each time period is used to represent the trend of available parking spaces in a parking lot in a group of parking lots during the target period of each time period. It can be the parking lot available parking space change value as shown in Table 4, or the parking lot available parking space quantity increase / decrease indicator as shown in Table 5 or Table 6. In this embodiment, there is no limitation on the available parking space change sequence corresponding to each time period.

[0128] Correspondingly, based on the characteristics of available parking spaces in each parking lot, a set of parking sequence characteristics for each parking lot is generated, including:

[0129] S41, determine the first trend quantity and the second trend quantity corresponding to each parking lot.

[0130] For the sequence of available parking spaces corresponding to each time period, the trend of available parking spaces in each parking lot within each time period can be determined, thereby obtaining the number of first trends and the number of second trends for each parking lot. The number of first trends for each parking lot is the number of parking lots whose trend is increasing, as indicated by the sequence of available parking spaces corresponding to each time period. The number of second trends for each parking lot is the number of parking lots whose trend is decreasing, as indicated by the sequence of available parking spaces corresponding to each time period.

[0131] S42, from a set of parking lots, each target parking lot and the corresponding change trend are determined sequentially to obtain the parking sequence characteristics of a set of parking lots.

[0132] Based on the number of first trends and the number of second trends corresponding to each parking lot, each target parking lot can be identified from a set of parking lots. Furthermore, the changing trend corresponding to each target parking lot can be determined, thus obtaining the parking sequence characteristics of the set of parking lots. Here, the parking sequence characteristics of the set of parking lots are the sequence obtained by arranging the changing trends corresponding to each target parking lot in a predetermined order. Specifically, one of the number of first trends and the number of second trends corresponding to each target parking lot is greater than a first preset number, and the changing trend corresponding to each target parking lot is the changing trend corresponding to the number of trends in its first and second trends that are greater than the first preset number.

[0133] For example, the sequence of changes in available parking spaces can be as shown in Table 7. Parking space information can be predicted based on sequence pattern mining of parking lot data; here, the sequence pattern can be parking sequence features. The prefixspan algorithm can be used to mine the sequence patterns of parking lot vacancy changes, that is, to mine the frequent changes in parking lot vacancy patterns across different time intervals within a certain period.

[0134] When performing sequence pattern mining, the prefixspan algorithm can be used to mine frequent sequence patterns of various lengths that satisfy the minimum support threshold in the empty space change sequences of each parking lot at each time point within a certain period of time. At the same time, a multiple minimum support strategy is used, and the calculation method of minimum support is shown in formula (2).

[0135] min_=su×p (2)

[0136] Where n is the number of days in the data collection period, and a is the minimum support rate (minimum support degree). The minimum support rate parameter can be adjusted according to the number of samples.

[0137] For example, the Prefixspan algorithm operates as follows:

[0138] Step 1: Find the prefix sum and corresponding projection dataset of the parking lot space increase / decrease sequence with a unit length of 1;

[0139] Step 2: Count the frequency of prefixes in the parking lot vacancy increase / decrease sequence and add prefixes with support higher than the minimum support threshold to the dataset to obtain frequent item set sequence patterns;

[0140] Step 3: Recursively mine all prefixes of length i that satisfy the minimum support requirement:

[0141] 1) Mine the projected dataset of the prefix; if the projected data is an empty set, return recursively.

[0142] 2) Calculate the minimum support of each item in the corresponding projected dataset, merge the items that meet the support requirement with the current prefix to obtain a new prefix, and recursively return if the support requirement is not met.

[0143] 3) Let i = i + 1, and let the prefixes be the new prefixes after merging individual terms. Recursively execute step 3 for each term.

[0144] Step 4: Return all frequent increase / decrease sequence patterns for each route in the parking lot space increase / decrease sequence sample set.

[0145] Based on the above Prefixspan algorithm operation steps, the following example illustrates the method of mining the time and parking space vacancy change trend sequence matrix.

[0146] First, a sequence of parking space availability changes for the same route at different periods can be constructed. This sequence includes changing road segment identifiers; for example, an increase in parking spaces at parking lot A is represented as "A+". The results are shown in Table 10.

[0147] Table 10

[0148] Route 1 20191201 8-9 o'clock A increases - B increases - C increases - D decreases - E increases - F decreases Route 1 20191202 8-9 o'clock A decreases - B increases - C increases - D decreases - E decreases - F decreases

[0149] Then, based on the prefixspan algorithm, the sequence patterns contained in the parking lot vacancy sequences of the route at the same time interval on different days are mined. Assuming that the minimum support threshold is set to 0.6, the frequency of all types of features is first counted, as shown in Table 11.

[0150] Table 11

[0151] frequency 1 1 2 2 2 1 1 2

[0152] Table 12 shows the prefixes and corresponding suffixes that satisfy the minimum support threshold:

[0153] Table 12

[0154]

[0155] Similarly, the binomial prefixes and corresponding suffixes that satisfy the minimum support threshold are shown in Table 13:

[0156] Table 13

[0157]

[0158] Table 14 shows the three prefixes and corresponding suffixes that satisfy the minimum support threshold:

[0159] Table 14

[0160]

[0161] Table 15 shows the four prefixes and corresponding suffixes that satisfy the minimum support threshold:

[0162] Table 15

[0163] B increases - C increases - D decreases - F decreases

[0164] The longest prefix sequence mined is used as the frequent sequence pattern of available parking spaces in each parking lot of the route within the time interval. This pattern will be updated and changed in real time with the data in the historical time range, and the latest trend will be mined in real time.

[0165] By using the embodiments provided in this application, and by statistically analyzing the increasing and decreasing trends of available parking spaces in each parking lot within the same time period at different time cycles, and determining parking sequence characteristics (i.e., frequent sequence patterns) based on the statistical results, the accuracy and convenience of determining the changing patterns of available parking spaces in parking lots can be improved.

[0166] As an optional approach, the traffic flow status characteristics of each navigation route include a traffic flow change sequence corresponding to each navigation route and each time period of multiple time periods. Each element in the traffic flow change sequence corresponding to each navigation route and each time period is used to represent the trend of traffic flow of one of the multiple road segments in the target time period of each time period relative to the traffic flow in the previous time period of the target time period.

[0167] For example, the traffic flow of the five road segments included in Route 1 at different times can be shown in Table 16.

[0168] Table 16

[0169]

[0170] Table 16 (continued)

[0171]

[0172] Each time period refers to a time interval; for example, 0 o'clock refers to the time interval from 0:00 to 1:00.

[0173] Correspondingly, based on the traffic flow state characteristics of each navigation route, traffic sequence characteristics for each navigation route are generated, including:

[0174] S51, determine the number of third trends corresponding to each segment among multiple segments of each navigation route and the number of fourth trends corresponding to each segment of each navigation route, wherein the number of third trends corresponding to each segment of each navigation route is the number of segments with an increasing trend in the changing trends of each navigation route indicated by the traffic flow change sequence corresponding to each navigation route and each time period, and the number of fourth trends corresponding to each segment of each navigation route is the number of segments with a decreasing trend in the changing trends of each navigation route indicated by the traffic flow change sequence corresponding to each navigation route and each time period;

[0175] S52, each target road segment and the corresponding change trend are sequentially determined from multiple road segments of each navigation route to obtain the traffic sequence characteristics of each navigation route. Among the number of third trends and the number of fourth trends corresponding to each target road segment of each navigation route, the number of trends corresponding to the change trend of each target road segment is greater than the second preset number.

[0176] For the traffic flow state characteristics of each navigation route, a similar method to generating a set of parking sequence characteristics for each parking lot based on the available parking spaces of each parking lot can be used. This method of generating a set of parking sequence characteristics for each navigation route based on the traffic flow state characteristics of each navigation route, without contradiction, can also be used in the process of generating traffic sequence characteristics for each navigation route based on the traffic flow state characteristics of each navigation route. This has already been explained and will not be repeated here.

[0177] By using the embodiments provided in this application, the accuracy and convenience of determining the traffic flow change patterns can be improved by statistically analyzing the increasing and decreasing trends of traffic flow in each segment of each navigation route within the same time period at different time cycles, and by determining the traffic sequence characteristics (i.e., frequent sequence patterns) of each navigation route based on the statistical results.

[0178] As an alternative approach, parking sequence features, traffic sequence features for each navigation route, and a set of first fusion features are input into the feature extraction layer of the target neural network for feature extraction, resulting in a set of second fusion features output by the feature extraction layer, including:

[0179] S61, input the parking sequence features, the traffic sequence features of each navigation route, and a set of first fusion features into the gated recurrent unit layer in the target neural network layer to obtain a set of second fusion features output by the gated recurrent unit layer; or,

[0180] S62, the parking sequence features, the traffic sequence features of each navigation route and a set of first fusion features are input into the multi-layer feedforward neural network layer in the target neural network layer to obtain a set of second fusion features output by the last feedforward neural network layer of the multi-layer feedforward neural network layer.

[0181] In this embodiment, various methods can be used to extract features from the traffic sequence features, parking sequence features, and a set of first fused features for each navigation route. These methods may include, but are not limited to, one of the following: feature extraction based on GRU (Gate Recurrent Unit), or feature extraction based on a multi-layer feedforward neural network. Correspondingly, the feature extraction layer may include, but is not limited to, one of the following: a GRU layer, or a multi-layer feedforward neural network (i.e., feedforward layers).

[0182] As an alternative implementation, feature extraction can be performed based on GRU, considering the parking sequence features, traffic sequence features for each navigation route, and a first set of fused features: the parking sequence features, traffic sequence features for each navigation route, and a first set of fused features are input into the GRU layer to obtain a second set of fused features output by the GRU layer. Compared to LSTM (Long Short-Term Memory), GRU is a model with fewer parameters that can handle sequence information well and can perform deep feature extraction.

[0183] For example, such as Figure 6 As shown, the traffic sequence features, parking sequence features, and a set of fused features output by the Transformer model for each navigation route can be input into the GRU layer to obtain the result of deep feature extraction by GRU. The input to the Transformer model is the available parking space features of each parking lot and the traffic flow state features of each navigation route. The available parking space features of all parking lots can be input as a whole, either once or separately for each navigation route's traffic flow state features.

[0184] As another alternative implementation, parking sequence features, traffic sequence features for each navigation route, and a set of first fusion features can be input into a multi-layer feedforward neural network layer to obtain a set of second fusion features output by the last feedforward neural network layer. At least some of the feedforward neural network layers in this multi-layer feedforward neural network layer are similar to GRU, that is, the GRU layer can be omitted and replaced with a multi-layer feedforward neural network layer, which can also effectively process and fuse features.

[0185] For example, such as Figure 7 As shown, the traffic sequence features, parking sequence features, and a set of fused features output by the Transformer model for each navigation route can be input into multiple feed forward layers to obtain the results of deep feature extraction by multiple feed forward layers.

[0186] The embodiments provided in this application demonstrate that deep feature extraction of fused features can be performed using multiple networks, thereby improving the flexibility of feature extraction.

[0187] As an optional approach, after inputting the parking sequence features, the traffic sequence features of each navigation route, and a set of first fusion features into the gated recurrent unit layer in the target neural network layer to obtain a set of second fusion features output by the gated recurrent unit layer, the above method further includes:

[0188] S71, a set of second fusion features is input into the feedforward neural network layer in the target neural network layer to obtain an updated set of second fusion features output by the feedforward neural network layer.

[0189] After obtaining a set of second fusion features, the effective information of other features can be processed by a feedforward neural network. This set of second fusion features can be output by a GRU layer, a multi-layer feedforward neural network, or other network layers; this embodiment does not limit this. After inputting the set of second fusion features into the feedforward neural network layer, an updated set of second fusion features output by the feedforward neural network layer can be obtained. The updated set of second fusion features can then be input into the aforementioned output layer to obtain the target parking path output by the output layer.

[0190] For example, such as Figure 8As shown, the traffic sequence features, parking sequence features, and a set of fused features output by the Transformer model for each navigation route can be input into the GRU layer. Next, the fused features are input into a feedforward neural network. The output of the feedforward neural network can be input into the output layer (e.g., the sigmoid function) to obtain the parking path output by the output layer. Here, parking route planning can be treated as a classification probability problem involving multiple paths. The network model constructed with the parking path classification probability as the output is as follows... Figure 8 As shown, the final output is the probability of parking route classification.

[0191] The embodiments provided in this application can improve the rationality of path selection by extracting effective information from other features through a feedforward neural network layer.

[0192] As an alternative approach, the characteristics of available parking spaces in each parking lot within a set of parking lots are determined, including:

[0193] S81, determine the change value of available parking spaces for each parking lot in each time period of multiple time periods, wherein the change value of available parking spaces for each parking lot in each time period is the difference between the number of available parking spaces at the end of the target period of each time period and the number of available parking spaces at the beginning of the target period of each time period.

[0194] S82, based on the change in available parking spaces for each parking lot in each time period, generate a sequence of available parking space changes corresponding to each time period, wherein each element in the sequence of available parking space changes corresponding to each time period is the change in available parking spaces for one parking lot in a group of parking lots in each time period.

[0195] In this embodiment, the vacancy characteristics of each parking lot can be a sequence of vacancy changes corresponding to each time period. Here, each element in the sequence of vacancy changes corresponding to each time period is the vacancy change value of the corresponding parking lot in each time period of a group of parking lots. The vacancy change value of each parking lot in each time period can be the difference between the number of vacancy spaces at the end of the target period of each time period and the number of vacancy spaces at the beginning of the target period of each time period, as shown in Table 4.

[0196] When determining the characteristics of available parking spaces in each parking lot, we can first determine the change value of available parking spaces in each time period of multiple time periods; then, sort the change value of available parking spaces in each time period according to the target parking space order to obtain the sequence of available parking space changes corresponding to each time period.

[0197] For example, as shown in Table 4, when the target time period is 0-1, the corresponding sequence of available parking spaces is -5-42-51-51-(-121)-24.

[0198] When fusing features of available parking spaces for each parking lot and traffic flow status features for each navigation route, the acquired sequence of changes in available parking spaces can be feature-encoded, for example, using one-hot encoding to obtain an available parking space encoding sequence corresponding to each time period. Furthermore, when fusing features of available parking spaces for each parking lot and traffic flow status features for each navigation route, the available parking space encoding sequences corresponding to each time period can be concatenated in chronological order to serve as the feature fusion characteristic for each parking lot.

[0199] Here, the result of feature encoding is the available parking space encoding sequence corresponding to each time period. Each parking lot corresponds to two elements in the available parking space encoding sequence corresponding to each time period. One element is used to represent the increasing trend of available parking spaces in each parking lot during the target period of each time period, and the other element is used to represent the decreasing trend of available parking spaces in each parking lot during the target period of each time period.

[0200] For example, the sequence of available parking spaces for each time period can be one-hot encoded. Referring to Table 4, the sequence of available parking spaces for time 1-2 is 11-(-135)-205-(-109)-28-(-35), and its one-hot encoding result is 10-01-10-01-10-01. These codes are then concatenated in chronological order to form the feature code of the parking space change sequence.

[0201] The embodiments provided in this application determine the sequence of available parking spaces for each time period by the difference between the end and start times of the same period in different time periods of each parking lot, and use the sequence of available parking spaces as the feature of available parking spaces. This can improve the convenience of determining the feature of available parking spaces and improve the rationality of the representation of the feature of available parking spaces.

[0202] As an alternative approach, traffic flow state characteristics for each navigation route in a set of navigation routes are determined, including:

[0203] S91, obtain the traffic flow sequence of each navigation route in each time period of multiple time periods, wherein each navigation route includes multiple road segments, and each element in the traffic flow sequence of each navigation route in each time period is the traffic flow of one of the multiple road segments in the target time period of each time period.

[0204] S92, obtain the traffic flow change sequence of each navigation route in each time period, wherein each element in the traffic flow change sequence of each navigation route in each time period is the difference between the traffic flow of one of the multiple road segments in the target time period and the traffic flow in the previous time period of the target time period.

[0205] S93, the traffic flow sequence of each navigation route in each time period, the traffic flow change sequence of each navigation route in each time period, and a set of preset features are concatenated into a traffic flow state feature vector of each navigation route in each time period, wherein the set of preset features is a set of features related to traffic flow.

[0206] For each navigation route, which may have multiple road segments, the traffic flow change sequence for each navigation route in each time period can be generated in a manner similar to that described above for generating the change sequence of available parking spaces corresponding to each time period. This has already been explained and will not be repeated here.

[0207] After obtaining the traffic flow change sequence of each navigation route in each time period, the traffic flow sequence of each navigation route in each time period, the traffic flow change sequence of each navigation route in each time period, and a set of preset features can be concatenated to obtain the traffic flow state feature vector of each navigation route in each time period. Here, the set of preset features is a set of features related to traffic flow, which may include, but is not limited to, at least one of the following: whether it is rush hour, whether it is a holiday, whether there is a large shopping area nearby, whether there are tourist attractions nearby, whether there is a subway or bus nearby, whether there are schools or hospitals nearby, etc. It may also include other factors related to population density, which are not limited in this embodiment.

[0208] For example, traffic flow data from various parking lots along a route (as shown in Table 16) is used to construct feature vectors for traffic flow at different time intervals within the same period. These feature vectors include the following information obtained at each time interval within the past period: traffic flow, changes in traffic flow compared to the previous moment, and other features. These other features include, but are not limited to, factors related to pedestrian density such as whether it is rush hour, whether it is a holiday, whether there are large shopping areas nearby, whether there are tourist attractions nearby, whether there are subways or buses nearby, and whether there are schools or hospitals nearby. The resulting traffic congestion feature vectors for each route are constructed, for example, [15,243,241,313,84,474,-5,42,51,51,-121,24,0,1,0,1…] (corresponding to times 1-2 in Table 16). Since the feature vectors for each time interval within the period are dynamically changing, there are multiple feature vectors for the same route.

[0209] The embodiments provided in this application construct a traffic flow change sequence for a navigation route within a given time period based on traffic flow, traffic flow changes, and preset features related to pedestrian density for different road segments within the same period. This can improve the rationality of the traffic flow change sequence construction.

[0210] As an optional approach, a set of parking lots corresponding to the navigation destination is identified, including:

[0211] S101, parking lots whose distance from the navigation destination is less than or equal to a preset distance threshold are identified as a group of parking lots.

[0212] In this embodiment, a distance threshold can be preset for selecting parking lots. When determining parking lots, all parking lots whose distance from the navigation endpoint is less than or equal to the preset distance threshold can be identified as a group of parking lots, using the navigation endpoint as a center point. Here, the parking lots whose distance from the navigation endpoint is less than or equal to the preset distance threshold can be either the parking lot whose center point is less than or equal to the preset distance threshold, or the parking lot whose farthest point is less than or equal to the preset distance threshold (the parking lot is located within a circle with the navigation endpoint as the center point and the preset distance threshold as the radius). The farthest point is the point within the parking lot that is furthest from the navigation endpoint.

[0213] The embodiments provided in this application demonstrate that by setting an appropriate distance threshold for selecting candidate parking lots, the rationality of vehicle navigation can be improved.

[0214] The following explanation, using optional examples, illustrates the navigation route processing method in this application embodiment. In this optional example, the available parking space feature of each parking lot in a group of parking lots is a sequence of available parking space changes corresponding to each time period; the traffic flow state feature of each navigation route is a traffic flow state feature vector of each navigation route in each time period; the parking sequence feature and traffic sequence feature are the mined sequence patterns; the Transformer model is used for feature fusion; the GRU layer is used for deep feature extraction; and the feedforward neural network layer is used to extract effective information from other features.

[0215] Intelligent traffic management and control can be applied to traffic route planning and navigation, map route navigation, and urban resource allocation, with parking space prediction and traffic flow status recognition being fundamental. Current technologies and methods for predicting parking spaces and congestion in urban traffic networks mainly include prediction models based on deep learning networks, prediction models based on clustering and genetic algorithms, and traffic congestion index prediction models based on nearest neighbors. However, all of these parking space prediction and congestion prediction schemes suffer from high prediction limitations and low accuracy.

[0216] This optional example provides a smart parking solution that integrates sequence pattern mining and Transformer. It uses sequence pattern mining to construct a Transformer to obtain parking solutions by fusing parking space changes and traffic congestion information from various time series. First, it acquires and standardizes road traffic data and parking lot data. Then, it mines parking lot data based on sequence patterns to predict parking space information for a specific time period. Next, it constructs a feature vector of traffic flow congestion data and mines the correlation between traffic and parking sequences using Transformer. Finally, it constructs a neural network to fuse parking and traffic information to obtain the optimal parking solution. This optional example constructs a smart parking strategy that combines time series traffic flow changes to predict the entire process from traffic to parking, thereby assisting traffic navigation. It can be applied to areas such as traffic route planning, urban resource allocation, and scientific map navigation.

[0217] Combination such as Figure 9 The process of processing navigation routes in this optional example may include the following steps:

[0218] Step S902: Acquire and standardize road traffic data and parking lot data.

[0219] Obtain road traffic and parking data at various time intervals (e.g., every hour) within the same time period (e.g., the same day), and standardize the road traffic and parking data. The standardization process may include the following steps:

[0220] Based on time interval settings, analyze the changing trends of parking lot vacancy spaces and traffic flow trends;

[0221] The increasing and decreasing trends (e.g., positive and negative differences in traffic flow within a time interval, positive and negative differences in available parking spaces) are marked with symbols, thereby obtaining the changing trends of parking space availability and traffic flow through the symbols.

[0222] Based on the changing trends of each parking lot along the route within time intervals, a sequence of changes in available parking spaces for each parking lot and a sequence of changes in traffic flow for each segment of each route are constructed.

[0223] Here, by constructing time intervals to analyze the trends of parking space changes and traffic flow increases and decreases, and mining the sequence patterns of traffic flow changes based on the trends, we can better perform real-time analysis of parking space changes and traffic flow changes and discover the patterns of change. At the same time, the analysis of parking space traffic flow changes has the advantage of real-time performance and is more efficient.

[0224] Step S904: Predict parking space information based on sequence pattern mining of parking lot data.

[0225] The prefixspan algorithm is used to mine sequence patterns of parking lot vacancy changes. This involves identifying frequent patterns of parking lot vacancy changes across different time intervals within a given period, resulting in parking sequence features. Similarly, sequence features of traffic flow changes along different routes can be mined to obtain traffic sequence features for each route.

[0226] Step S906: Construct a feature vector for traffic flow congestion data.

[0227] Traffic flow congestion data feature vectors are constructed based on traffic flow, comparison with the traffic flow changes of the previous moment, and other features. Since the feature vectors of each time interval within the cycle are dynamically changing, there are multiple feature vectors for the same route.

[0228] Step S908: Mining the correlation between traffic sequences and parking sequences based on Transformer.

[0229] There is a certain correlation and causal relationship between traffic flow change sequences and parking space change sequences. The correlation information between traffic sequences and parking sequences can be mined using the Transformer model. One-hot encoding can be performed on the elements of the parking space availability sequence, and the codes can be concatenated in chronological order to serve as the feature encoding for the parking space change sequence. Then, the feature encoding of the parking space change sequence and the feature vector of traffic flow congestion data are input into the Transformer model to obtain two feature representations.

[0230] Here, by mining the correlation between traffic flow sequences and parking space change sequences using Transformer, we can better uncover the relationships between parking scheme selection features, extract deeper features, and thus improve the accuracy of parking scheme planning.

[0231] Step S910: Construct a parking solution that integrates parking and traffic information using a neural network.

[0232] The parking sequence features, traffic sequence features, and the output of the Transformer model are input into the GRU layer for feature extraction. The extracted features are then input into the feedforward neural network to process the effective information of other features. Finally, the output of the feedforward neural network is processed using the sigmoid function as the output layer and the cross-entropy loss as the loss function, and the parking path classification probability is finally output.

[0233] This optional example demonstrates how real-time prediction of parking space information (parking space change patterns) and traffic flow feature vectors through sequence patterns enables better real-time analysis of available parking space changes. By fusing parking space change patterns and traffic flow features to construct a Transformer for deeper mining of relevant features, the flexibility of sequence pattern mining and the generalization ability of the model can be improved, the impact of noise features can be reduced, and the accuracy of traffic congestion identification can be improved.

[0234] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0235] According to another aspect of the embodiments of this application, a navigation route processing apparatus for implementing the above-described navigation route processing method is also provided. Figure 10 This is a schematic diagram of the structure of an optional navigation route processing device according to an embodiment of this application, as shown below. Figure 10 As shown, the device may include:

[0236] The first determining unit 1002 is used to determine a set of navigation routes between the navigation starting point and the navigation ending point, and to determine a set of parking lots corresponding to the navigation ending point, thereby obtaining a set of parking paths. Each parking path in the set of parking paths includes a combination of a navigation route in the set of navigation routes and a parking lot in the set of parking lots.

[0237] The second determining unit 1004 is connected to the first determining unit 1002 and is used to determine the vacant parking space characteristics of each parking lot in a group of parking lots. The vacant parking space characteristics of each parking lot are used to describe the historical vacant parking spaces of each parking lot during the target time period, which is the time period from the navigation starting point to the navigation destination.

[0238] The third determining unit 1006, connected to the second determining unit 1004, is used to determine the traffic flow state characteristics of each navigation route in a set of navigation routes, wherein the traffic flow state characteristics of each navigation route are used to describe the historical traffic flow state of each navigation route corresponding to the target time period.

[0239] The first input unit 1008, connected to the third determining unit 1006, is used to input the features of available parking spaces in each parking lot and the traffic flow status features of each navigation route into the target prediction model to obtain the target parking path output by the target prediction model. The target prediction model is used to predict the selection probability of each parking path based on the correlation between the features of available parking spaces in each parking lot and the traffic flow status features of each navigation route.

[0240] It should be noted that the first determining unit 1002 in this embodiment can be used to execute the above step S202, the second determining unit 1004 in this embodiment can be used to execute the above step S204, the third determining unit 1006 in this embodiment can be used to execute the above step S206, and the first input unit 1008 in this embodiment can be used to execute the above step S208.

[0241] The embodiments provided in this application determine a set of navigation routes between a navigation start point and a navigation end point, and determine a set of parking lots corresponding to the navigation end point, thus obtaining a set of parking paths. Each parking path in the set of parking paths includes a combination of one navigation route from the set of navigation routes and one parking lot from the set of parking lots. The available parking space characteristics of each parking lot in the set of parking lots are determined, wherein the available parking space characteristics of each parking lot describe the historical available parking spaces of each parking lot during a target time period, where the target time period is the time from the navigation start point to the navigation end point. The traffic flow state characteristics of each navigation route in the set of navigation routes are also determined. The traffic flow state features of each navigation route are used to describe the historical traffic flow state of each navigation route corresponding to the target time period. The available parking space features of each parking lot and the traffic flow state features of each navigation route are input into the target prediction model to obtain the target parking path output by the target prediction model. The target prediction model is used to predict the selection probability of each parking path based on the correlation between the available parking space features of each parking lot and the traffic flow state features of each navigation route. This solves the problem of unreasonable navigation route planning caused by the poor parameter flexibility of path planning in the navigation route processing methods of related technologies, and improves the rationality of navigation route planning.

[0242] As an optional solution, the first input unit includes:

[0243] The first input module is used to input the available parking space features of each parking lot and the traffic flow status features of each navigation route into the feature fusion layer of the target prediction model to obtain a set of first fusion features output by the feature fusion layer.

[0244] The second input module is used to input the available parking space features of each parking lot, the traffic flow status features of each navigation route, and a set of first fusion features into the target neural network layer of the target prediction model to obtain the selection probability of each parking path output by the target neural network layer.

[0245] The first determining module is used to determine the parking path with the highest selection probability from a set of parking paths as the target parking path.

[0246] Optional examples of this implementation scheme can be found in the examples shown in the above navigation route processing method, which will not be repeated here.

[0247] As an optional solution, the first input module includes:

[0248] The first input submodule is used to input the available parking space features of each parking lot and the traffic flow status features of each navigation route into the encoder model of the target prediction model, and obtain a set of first fusion features output by the encoder model.

[0249] Optional examples of this implementation scheme can be found in the examples shown in the above navigation route processing method, which will not be repeated here.

[0250] As an optional solution, the second input module includes:

[0251] The first generation submodule is used to generate a set of parking sequence features for a parking lot based on the characteristics of the available parking spaces of each parking lot. The parking sequence features are used to represent a set of target parking lots in a set of parking lots, and the changing trend of available parking spaces of each target parking lot in a set of target parking lots. A set of target parking lots is a set of parking lots in which the number of parking lots with the same changing trend of available parking spaces is greater than a first preset number within a target time period of multiple time periods. The changing trend of available parking spaces of each target parking lot is the changing trend of the number of available parking spaces of each target parking lot in a target time period of multiple time periods being greater than the first preset number.

[0252] The second generation submodule is used to generate traffic sequence features for each navigation route based on the traffic flow status features of each navigation route. Each navigation route includes multiple road segments, and the traffic sequence features of each navigation route are used to represent a set of target road segments for each navigation route, as well as the traffic flow change trend of each target road segment in the set of target road segments. A set of target road segments is a set of multiple road segments of each navigation route in which the number of road segments with the same traffic flow change trend is greater than a second preset number in the target time period of multiple time periods. The traffic flow change trend of each target road segment is the traffic flow change trend of each target road segment in which the number of road segments with the same traffic flow change trend is greater than a second preset number in the target time period of multiple time periods.

[0253] The second input submodule is used to input the parking sequence features, the traffic sequence features of each navigation route and a set of first fusion features into the feature extraction layer in the target neural network layer for feature extraction, and obtain a set of second fusion features output by the feature extraction layer.

[0254] The third input submodule is used to input a set of second fusion features into the output layer of the target neural network layer to obtain the output result of the output layer, wherein the output result of the output layer is used to represent the selection probability of each parking path.

[0255] Optional examples of this implementation scheme can be found in the examples shown in the above navigation route processing method, which will not be repeated here.

[0256] As an optional approach, the available parking space characteristics of each parking lot include a sequence of available parking space changes corresponding to each of multiple time periods. Each element in the sequence of available parking space changes corresponding to each time period represents the trend of available parking spaces changing in one parking lot within a group of parking lots during a target time period of each time period. Correspondingly, the first generation submodule includes:

[0257] The first determining subunit is used to determine the first trend quantity and the second trend quantity corresponding to each parking lot. The first trend quantity corresponding to each parking lot is the number of parking lot changes with an increasing trend indicated by the sequence of available parking spaces changes corresponding to each time period. The second trend quantity corresponding to each parking lot is the number of parking lot changes with a decreasing trend indicated by the sequence of available parking spaces changes corresponding to each time period.

[0258] The second determining subunit is used to sequentially determine each target parking lot and the changing trend corresponding to each target parking lot from a set of parking lots, thereby obtaining the parking sequence characteristics of a set of parking lots. Among the first trend number and the corresponding second trend number for each target parking lot, the trend number corresponding to the changing trend of each target parking lot is greater than the first preset number.

[0259] Optional examples of this implementation scheme can be found in the examples shown in the above navigation route processing method, which will not be repeated here.

[0260] As an optional approach, the traffic flow status characteristics of each navigation route include a traffic flow change sequence corresponding to each navigation route and each time period across multiple time periods. Each element in the traffic flow change sequence corresponding to each navigation route and each time period represents the trend of traffic flow in one of the multiple road segments during the target time period relative to the traffic flow in the previous time period. Correspondingly, the second generation submodule includes:

[0261] The third determining subunit is used to determine the number of third trends corresponding to each of the multiple road segments of each navigation route and the number of fourth trends corresponding to each road segment of each navigation route. The number of third trends corresponding to each road segment of each navigation route is the number of road segments of each navigation route whose trend is an increasing trend, as indicated by the traffic flow change sequence corresponding to each navigation route and each time period. The number of fourth trends corresponding to each road segment of each navigation route is the number of road segments of each navigation route whose trend is a decreasing trend, as indicated by the traffic flow change sequence corresponding to each navigation route and each time period.

[0262] The fourth determining subunit is used to sequentially determine each target road segment and the corresponding change trend from multiple road segments of each navigation route to obtain the traffic sequence characteristics of each navigation route. Among the number of third trends and the number of fourth trends corresponding to each target road segment of each navigation route, the number of trends corresponding to the change trend of each target road segment is greater than the second preset number.

[0263] Optional examples of this implementation scheme can be found in the examples shown in the above navigation route processing method, which will not be repeated here.

[0264] As an optional solution, the second input submodule includes:

[0265] The first input subunit is used to input the parking sequence features, the traffic sequence features of each navigation route, and a set of first fusion features into the gated recurrent unit layer in the target neural network layer, to obtain a set of second fusion features output by the gated recurrent unit layer; or,

[0266] The second input subunit is used to input the parking sequence features, the traffic sequence features of each navigation route, and a set of first fusion features into the multi-layer feedforward neural network layer in the target neural network layer, so as to obtain a set of second fusion features output by the last feedforward neural network layer of the multi-layer feedforward neural network layer.

[0267] Optional examples of this implementation scheme can be found in the examples shown in the above navigation route processing method, which will not be repeated here.

[0268] As an optional solution, the above-mentioned device further includes:

[0269] The second input unit is used to input the parking sequence features, the traffic sequence features of each navigation route and a set of first fusion features into the gated recurrent unit layer in the target neural network layer to obtain a set of second fusion features output by the gated recurrent unit layer, and then input a set of second fusion features into the feedforward neural network layer in the target neural network layer to obtain an updated set of second fusion features output by the feedforward neural network layer.

[0270] Optional examples of this implementation scheme can be found in the examples shown in the above navigation route processing method, which will not be repeated here.

[0271] As an optional solution, the second determining unit includes:

[0272] The second determining module is used to determine the change value of available parking spaces for each parking lot in each time period of multiple time periods, wherein the change value of available parking spaces for each parking lot in each time period is the difference between the number of available parking spaces at the end of the target period of each time period and the number of available parking spaces at the beginning of the target period of each time period.

[0273] The generation module is used to generate a sequence of available parking space changes corresponding to each time period based on the change value of available parking spaces in each parking lot in each time period. Each element in the sequence of available parking space changes corresponding to each time period is the change value of available parking spaces in one parking lot in a group of parking lots in each time period.

[0274] Optional examples of this implementation scheme can be found in the examples shown in the above navigation route processing method, which will not be repeated here.

[0275] As an optional approach, the third determining unit includes:

[0276] The first acquisition module is used to acquire the traffic flow sequence of each navigation route in each time period of multiple time periods. Each navigation route includes multiple road segments, and each element in the traffic flow sequence of each navigation route in each time period is the traffic flow of one of the multiple road segments in the target time period of each time period.

[0277] The second acquisition module is used to acquire the traffic flow change sequence of each navigation route in each time period. Each element in the traffic flow change sequence of each navigation route in each time period is the difference between the traffic flow of one of the multiple road segments in the target time period and the traffic flow in the previous time period of the target time period.

[0278] The splicing module is used to splice the traffic flow sequence of each navigation route in each time period, the traffic flow change sequence of each navigation route in each time period, and a set of preset features into a traffic flow state feature vector of each navigation route in each time period. The set of preset features is a set of features associated with traffic flow.

[0279] Optional examples of this implementation scheme can be found in the examples shown in the above navigation route processing method, which will not be repeated here.

[0280] As an optional approach, the first determining unit includes:

[0281] The third determination module is used to determine parking lots whose distance from the navigation destination is less than or equal to a preset distance threshold as a group of parking lots.

[0282] Optional examples of this implementation scheme can be found in the examples shown in the above navigation route processing method, which will not be repeated here.

[0283] According to another aspect of the embodiments of this application, an electronic device for implementing the above-described navigation route processing method is also provided. This electronic device may be... Figure 1 The terminal device or server shown. This embodiment uses this electronic device as an example for illustration. Figure 11 As shown, the electronic device includes a memory 1102 and a processor 1104. The memory 1102 stores a computer program, and the processor 1104 is configured to execute the steps of any of the above method embodiments via the computer program.

[0284] Optionally, in this embodiment, the aforementioned electronic device may be located in at least one of a plurality of network devices in a computer network.

[0285] Optionally, in this embodiment, the processor can be configured to perform the following steps via a computer program:

[0286] S1, determine a set of navigation routes between the navigation start point and the navigation end point, and determine a set of parking lots corresponding to the navigation end point, to obtain a set of parking paths, wherein each parking path in the set of parking paths includes a combination of a navigation route in the set of navigation routes and a parking lot in the set of parking lots;

[0287] S2, determine the available parking space characteristics of each parking lot in a set of parking lots, wherein the available parking space characteristics of each parking lot are used to describe the historical available parking spaces of each parking lot during the target time period, which is the time period from the navigation starting point to the navigation destination.

[0288] S3, determine the traffic flow state characteristics of each navigation route in a set of navigation routes, wherein the traffic flow state characteristics of each navigation route are used to describe the historical traffic flow state of each navigation route corresponding to the target time period;

[0289] S4. Input the available parking space features of each parking lot and the traffic flow status features of each navigation route into the target prediction model to obtain the target parking path output by the target prediction model. The target prediction model is used to predict the selection probability of each parking path based on the correlation between the available parking space features of each parking lot and the traffic flow status features of each navigation route.

[0290] Alternatively, as those skilled in the art will understand, Figure 11 The structure shown is for illustrative purposes only. Electronic devices can also be smartphones (such as Android phones, iOS phones, etc.), tablets, PDAs, and other terminal devices such as MIDs and PADs. Figure 11 This does not limit the structure of the aforementioned electronic devices. For example, the electronic device may also include components that are more... Figure 11 The more or fewer components shown (such as network interfaces, etc.), or having the same Figure 11 The different configurations shown.

[0291] The memory 1102 can be used to store software programs and modules, such as the program instructions / modules corresponding to the navigation route processing method and apparatus in this embodiment. The processor 1104 executes various functional applications and data processing by running the software programs and modules stored in the memory 1102, thereby implementing the aforementioned navigation route processing method. The memory 1102 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 1102 may further include memory remotely located relative to the processor 1104, and these remote memories can be connected to the terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof. Specifically, the memory 1102 may be used, but is not limited to, for serializing files and compiling files, etc.

[0292] As an example, such as Figure 11 As shown, the memory 1102 may include, but is not limited to, the first determining unit 1002, the second determining unit 1004, the third determining unit 1006, and the first input unit 1008 from the navigation route processing device. Furthermore, it may include, but is not limited to, other module units from the navigation route processing device, which will not be elaborated upon in this example.

[0293] Optionally, the transmission device 1106 described above is used to receive or send data via a network. Specific examples of the network described above may include wired networks and wireless networks. In one example, the transmission device 1106 includes a Network Interface Controller (NIC), which can be connected to other network devices and a router via a network cable to communicate with the Internet or a local area network. In another example, the transmission device 1106 is a radio frequency (RF) module, used for wireless communication with the Internet.

[0294] In addition, the aforementioned electronic device also includes: a display 1108 for displaying navigation routes, navigation start points, navigation destinations, etc.; and a connection bus 1110 for connecting various module components in the aforementioned electronic device.

[0295] In other embodiments, the aforementioned terminal device or server can be a node in a distributed system, wherein the distributed system can be a blockchain system, which is a distributed system formed by connecting multiple nodes through network communication. The nodes can form a peer-to-peer (P2P) network, and any form of computing device, such as a server, terminal, or other electronic device, can become a node in the blockchain system by joining this peer-to-peer network.

[0296] According to one aspect of this application, a computer program product is provided, comprising a computer program / instructions containing program code for performing the methods shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via communication section 1209, and / or installed from removable medium 1211. When the computer program is executed by central processing unit 1201, it performs various functions provided in the embodiments of this application. The above embodiment numbers are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0297] Figure 12 A schematic block diagram of a computer system architecture for implementing embodiments of the present application is shown. Figure 12As shown, the computer system 1200 includes a central processing unit (CPU) 1201, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 1202 or programs loaded from storage section 1208 into random access memory (RAM). The RAM 1203 also stores various programs and data required for system operation. The CPU 1201, ROM 1202, and RAM 1203 are interconnected via a bus 1204. An input / output interface 1205 (I / O interface) is also connected to the bus 1204.

[0298] The following components are connected to the input / output interface 1205: an input section 1206 including a keyboard, mouse, etc.; an output section 1207 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 1208 including a hard disk, etc.; and a communication section 1209 including a network interface card such as a local area network card, modem, etc. The communication section 1209 performs communication processing via a network such as the Internet. A drive 1210 is also connected to the input / output interface 1205 as needed. A removable medium 1211, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 1210 as needed so that computer programs read from it can be installed into the storage section 1208 as needed.

[0299] Specifically, according to embodiments of this application, the processes described in the various method flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 1209, and / or installed from removable medium 1211. When the computer program is executed by central processing unit 1201, it performs various functions defined in the system of this application.

[0300] It should be noted that, Figure 12 The computer system 1200 of the electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0301] According to one aspect of this application, a computer-readable storage medium is provided, from which a processor of a computer device reads computer instructions, and the processor executes the computer instructions, causing the computer device to perform the methods provided in various optional implementations of the above embodiments.

[0302] Optionally, in this embodiment, the computer-readable storage medium may be configured to store a computer program for performing the following steps:

[0303] S1, determine a set of navigation routes between the navigation start point and the navigation end point, and determine a set of parking lots corresponding to the navigation end point, to obtain a set of parking paths, wherein each parking path in the set of parking paths includes a combination of a navigation route in the set of navigation routes and a parking lot in the set of parking lots;

[0304] S2, determine the available parking space characteristics of each parking lot in a set of parking lots, wherein the available parking space characteristics of each parking lot are used to describe the historical available parking spaces of each parking lot during the target time period, which is the time period from the navigation starting point to the navigation destination.

[0305] S3, determine the traffic flow state characteristics of each navigation route in a set of navigation routes, wherein the traffic flow state characteristics of each navigation route are used to describe the historical traffic flow state of each navigation route corresponding to the target time period;

[0306] S4. Input the available parking space features of each parking lot and the traffic flow status features of each navigation route into the target prediction model to obtain the target parking path output by the target prediction model. The target prediction model is used to predict the selection probability of each parking path based on the correlation between the available parking space features of each parking lot and the traffic flow status features of each navigation route.

[0307] Optionally, in this embodiment, those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing the hardware related to the terminal device. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0308] If the integrated units in the above embodiments are implemented as software functional units and sold or used as independent products, they can be stored in the aforementioned computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause one or more computer devices (which may be personal computers, servers, or network devices, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application.

[0309] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0310] In the several embodiments provided in this application, it should be understood that the disclosed client can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, indirect coupling or communication connection between units or modules, and may be electrical or other forms.

[0311] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0312] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or at least two units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0313] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for processing navigation routes, characterized in that, include: A set of navigation routes between the navigation start point and the navigation end point is determined, and a set of parking lots corresponding to the navigation end point is determined to obtain a set of parking paths, wherein each parking path in the set of parking paths includes a combination of a navigation route in the set of navigation routes and a parking lot in the set of parking lots; Determine the available parking space characteristics of each parking lot in the group of parking lots to obtain a set of available parking space characteristics, wherein the available parking space characteristics of each parking lot are used to describe the historical available parking spaces of each parking lot during a target time period, the target time period being the time period from the navigation starting point to the navigation destination; Determine the traffic flow state characteristics of each navigation route in the set of navigation routes to obtain a set of traffic flow state characteristics, wherein the traffic flow state characteristics of each navigation route are used to describe the historical traffic flow state of each navigation route corresponding to the target time period; The set of available parking space features and the set of traffic flow state features are input into the feature fusion layer of the target prediction model to obtain a set of first fusion features. The target prediction model is used to predict the selection probability of each parking route based on the correlation between the available parking space features of each parking lot and the traffic flow state features of each navigation route. In the target neural network layer of the target prediction model, feature extraction is performed on a set of parking sequence features determined based on the set of available parking space features, a set of traffic sequence features determined based on the set of traffic flow state features, and the set of first fusion features to obtain a set of second fusion features. Based on the set of second fusion features, the selection probability of each parking path is determined, and based on the selection probability of each parking path, the target parking path is determined from the set of parking paths.

2. The method according to claim 1, characterized in that, Determining the target parking path from the set of parking paths based on the selection probability of each parking path includes: The parking path with the highest selection probability among the set of parking paths is determined as the target parking path.

3. The method according to claim 1, characterized in that, The step of inputting the set of available parking space features and the set of traffic flow state features into the feature fusion layer of the target prediction model to obtain a first set of fused features includes: The available parking space features of each parking lot and the traffic flow status features of each navigation route are input into the encoder model of the target prediction model to obtain the first set of fused features output by the encoder model.

4. The method according to claim 1, characterized in that, In the target neural network layer of the target prediction model, before extracting features from a set of parking sequence features determined based on the set of available parking space features, a set of traffic sequence features determined based on the set of traffic flow state features, and the set of first fusion features to obtain a set of second fusion features, the process includes: Based on the available parking space characteristics of each parking lot, a parking sequence characteristic of the group of parking lots is generated. The parking sequence characteristic is used to represent a group of target parking lots in the group of parking lots, and the available parking space change trend of each target parking lot in the group of target parking lots. The group of target parking lots is the group of parking lots in which the number of parking lots with the same available parking space change trend is greater than a first preset number within the target time period of multiple time periods. The available parking space change trend of each target parking lot is the change trend of the number of available parking spaces in each target parking lot being greater than the first preset number within the target time period of multiple time periods. Based on the traffic flow status characteristics of each navigation route, a traffic sequence characteristic is generated for each navigation route. Each navigation route includes multiple road segments. The traffic sequence characteristic of each navigation route is used to represent a set of target road segments for each navigation route, and the traffic flow change trend of each target road segment in the set of target road segments. The set of target road segments refers to the road segments in the multiple road segments of each navigation route where the number of road segments exhibiting the same traffic flow change trend is greater than a second preset number within the target time period of the multiple time periods. The traffic flow change trend of each target road segment is the number of traffic flow change trends that each target road segment exhibits within the target time period of the multiple time periods that is greater than the second preset number. The parking sequence features, the traffic sequence features of each navigation route, and the first set of fusion features are input into the feature extraction layer in the target neural network layer for feature extraction, resulting in a set of second fusion features output by the feature extraction layer. Based on the set of second fusion features, the selection probability of each parking path is determined, and according to the selection probability of each parking path, the following is included: The set of second fusion features is input into the output layer of the target neural network layer to obtain the output result of the output layer, wherein the output result of the output layer is used to represent the selection probability of each parking path.

5. The method according to claim 4, characterized in that, The vacancy characteristics of each parking lot include a sequence of vacancy changes corresponding to each of the plurality of time periods, wherein each element in the sequence of vacancy changes corresponding to each time period is used to represent the trend of vacancy changes of one parking lot in the group of parking lots during the target time period of each time period. The step of generating parking sequence features for the group of parking lots based on the available parking space characteristics of each parking lot includes: Determine the first trend quantity and the second trend quantity corresponding to each parking lot, wherein the first trend quantity corresponding to each parking lot is the number of parking lot changes in each parking lot whose trend is an increasing trend, as indicated by the parking space change sequence corresponding to each time period; and the second trend quantity corresponding to each parking lot is the number of parking lot changes in each parking lot whose trend is a decreasing trend, as indicated by the parking space change sequence corresponding to each time period. From the group of parking lots, each target parking lot and the corresponding change trend are determined sequentially to obtain the parking sequence characteristics of the group of parking lots. Among the first trend number and the corresponding second trend number for each target parking lot, the trend number corresponding to the change trend for each target parking lot is greater than the first preset number.

6. The method according to claim 4, characterized in that, The traffic flow status characteristics of each navigation route include a traffic flow change sequence corresponding to each navigation route and each of the multiple time periods, wherein each element in the traffic flow change sequence corresponding to each navigation route and each time period is used to represent the trend of traffic flow of one of the multiple road segments in the target time period of each time period relative to the traffic flow in the previous time period of the target time period; The step of generating traffic sequence features for each navigation route based on the traffic flow state features of each navigation route includes: Determine the number of third trends corresponding to each of the plurality of road segments of each navigation route and the number of fourth trends corresponding to each of the road segments of each navigation route, wherein the number of third trends corresponding to each of the road segments of each navigation route is the number of road segments of each navigation route whose change trend is an increasing trend, as indicated by the traffic flow change sequence corresponding to each navigation route and each time period; and the number of fourth trends corresponding to each of the road segments of each navigation route is the number of road segments of each navigation route whose change trend is a decreasing trend, as indicated by the traffic flow change sequence corresponding to each navigation route and each time period. Each target road segment and the corresponding change trend are sequentially determined from the multiple road segments of each navigation route to obtain the traffic sequence characteristics of each navigation route. Among the number of third trends and the number of fourth trends corresponding to each target road segment of each navigation route, the number of trends corresponding to the change trend of each target road segment is greater than the second preset number.

7. The method according to claim 1, characterized in that, In the target neural network layer of the target prediction model, feature extraction is performed on a set of parking sequence features determined based on the set of available parking space features, a set of traffic sequence features determined based on the set of traffic flow state features, and the set of first fusion features to obtain a set of second fusion features, including: The parking sequence features, the traffic sequence features of each navigation route, and the first set of fused features are input into a gated recurrent unit layer in the target neural network layer to obtain the second set of fused features output by the gated recurrent unit layer; or, The parking sequence features, the traffic sequence features of each navigation route, and the first set of fusion features are input into the multi-layer feedforward neural network layer in the target neural network layer to obtain the second set of fusion features output by the last feedforward neural network layer of the multi-layer feedforward neural network layer.

8. The method according to claim 7, characterized in that, After inputting the parking sequence features, the traffic sequence features of each navigation route, and the first set of fused features into the gated recurrent unit layer in the target neural network layer to obtain the second set of fused features output by the gated recurrent unit layer, the method further includes: The set of second fusion features is input into the feedforward neural network layer in the target neural network layer to obtain the updated set of second fusion features output by the feedforward neural network layer.

9. The method according to claim 1, characterized in that, Determining the characteristics of available parking spaces in each of the group of parking lots includes: Determine the change in available parking spaces for each parking lot in each of the multiple time periods, wherein the change in available parking spaces for each parking lot in each time period is the difference between the number of available parking spaces for each parking lot at the end of the target period in each time period and the number of available parking spaces for each parking lot at the beginning of the target period in each time period. Based on the change in available parking spaces for each parking lot in each time period, a sequence of available parking space changes corresponding to each time period is generated, wherein each element in the sequence of available parking space changes corresponding to each time period is the change in available parking spaces for one of the parking lots in the group of parking lots in each time period.

10. The method according to claim 1, characterized in that, Determining the traffic flow state characteristics of each navigation route in the set of navigation routes includes: Obtain the traffic flow sequence of each navigation route in each of the multiple time periods, wherein each navigation route includes multiple road segments, and each element of the traffic flow sequence of each navigation route in each time period is the traffic flow of one of the multiple road segments in the target time period of each time period; Obtain the traffic flow change sequence for each navigation route in each time period, wherein each element in the traffic flow change sequence for each navigation route in each time period is the difference between the traffic flow of one of the multiple road segments in the target time period and the traffic flow in the previous time period of the target time period. The traffic flow sequence of each navigation route in each time period, the traffic flow change sequence of each navigation route in each time period, and a set of preset features are concatenated to form the traffic flow state feature vector of each navigation route in each time period, wherein the set of preset features is a set of features associated with traffic flow.

11. The method according to any one of claims 1 to 10, characterized in that, The determination of a set of parking lots corresponding to the navigation destination includes: Parking lots whose distance from the navigation destination is less than or equal to a preset distance threshold are identified as the group of parking lots.

12. A navigation route processing device, characterized in that, include: The first determining unit is used to determine a set of navigation routes between the navigation starting point and the navigation ending point, and to determine a set of parking lots corresponding to the navigation ending point, thereby obtaining a set of parking paths. Each parking path in the set of parking paths includes a combination of a navigation route in the set of navigation routes and a parking lot in the set of parking lots. The second determining unit is used to determine the available parking space characteristics of each parking lot in the group of parking lots to obtain a set of available parking space characteristics, wherein the available parking space characteristics of each parking lot are used to describe the historical available parking spaces of each parking lot during a target time period, and the target time period is the time period from the navigation starting point to the navigation destination. The third determining unit is used to determine the traffic flow state characteristics of each navigation route in the set of navigation routes to obtain a set of traffic flow state characteristics, wherein the traffic flow state characteristics of each navigation route are used to describe the historical traffic flow state of each navigation route corresponding to the target time period. The first input unit is used to input the set of available parking space features and the set of traffic flow state features into the feature fusion layer of the target prediction model to obtain a set of first fusion features. The target prediction model is used to predict the selection probability of each parking path based on the correlation between the available parking space features of each parking lot and the traffic flow state features of each navigation route. In the target neural network layer of the target prediction model, feature extraction is performed on a set of parking sequence features determined based on the set of available parking space features, a set of traffic sequence features determined based on the set of traffic flow state features, and the set of first fusion features to obtain a set of second fusion features. The selection probability of each parking path is determined based on the set of second fusion features, and a target parking path is determined from the set of parking paths according to the selection probability of each parking path.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein the program, when executed, performs the method described in any one of claims 1 to 11.

14. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method according to any one of claims 1 to 11.

15. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to execute the method described in any one of claims 1 to 11 through the computer program.

Images