Travel time prediction method, travel time prediction model training method, device, and medium

Abstract

The present disclosure provides a travel time prediction method, a travel time prediction model training method, equipment and a medium, relates to the technical field of artificial intelligence, and in particular to the technical field of intelligent traffic, deep learning and map navigation. The specific implementation scheme is: constructing a bus line map and a bus station map according to road network data, bus network data and traffic condition data; obtaining a query bus route between a starting point and a terminal point, station heat information of stations in the query bus route and current time information; and predicting a bus travel time of the query bus route by using the bus line map, the bus station map, the query bus route, the station heat information and the current time information. Through the above technical scheme, the prediction accuracy of the bus travel time can be improved.

Description

Technical Field

[0001] This disclosure relates to the field of artificial intelligence technology, and in particular to the fields of intelligent transportation, deep learning, and map navigation technology. Background Technology

[0002] Public transport travel time information is a high-frequency, essential need for people's travel. It is widely used in frequent scenarios such as bus route planning, real-time bus updates, and electronic bus stop displays. Accurate prediction of bus travel time is of great value in improving user travel efficiency and product reputation.

[0003] Existing methods for estimating bus travel time generally employ a holistic strategy or unified model. They primarily utilize information such as road length, attributes, traffic lights, real-time traffic conditions, and historical traffic conditions along the predicted route for regression processing. However, different routes exhibit significant variations in road length, attributes, traffic lights, and historical traffic conditions, rendering existing feature processing methods inadequate in capturing spatiotemporal characteristics. Furthermore, current bus time estimation methods do not consider the topological characteristics of the road network. In addition, existing methods do not adequately utilize traffic condition and driving trajectory information. Therefore, a more effective method for predicting bus travel time is needed. Summary of the Invention

[0004] This disclosure provides a method for predicting travel time, a training method for a travel time prediction model, as well as an apparatus and medium.

[0005] According to one aspect of this disclosure, a method for predicting travel time is provided, the method comprising:

[0006] Based on road network data, public transport network data, and traffic condition data, construct a public transport route map and a public transport stop map;

[0007] Obtain the bus route between the origin and destination, the station popularity information of the bus route, and the current time information;

[0008] Using the bus route map, the bus stop map, the queried bus route, the stop popularity information, and the current time information, the bus travel time for the queried bus route is predicted.

[0009] According to another aspect of this disclosure, a method for training a travel time prediction model is provided, the method comprising:

[0010] Based on road network data, public transport network data, and traffic condition data, construct a public transport route map and a public transport stop map;

[0011] Obtain the bus route between the origin and destination, the station popularity information of the bus route, and the current time information;

[0012] Using the bus route map, the bus stop map, the queried bus route, the stop popularity information, and the current time information, the bus travel time of the queried bus route is predicted;

[0013] A travel time prediction model is trained based on the bus travel time and the first label data of the queried bus route.

[0014] According to another aspect of this disclosure, a travel time prediction device is provided, the device comprising:

[0015] The first diagram structure determination module is used to construct bus route maps and bus stop maps based on road network data, public transport network data, and traffic condition data.

[0016] The first data acquisition module is used to acquire the query bus route between the origin and the destination, the station popularity information of the stations in the query bus route, and the current time information;

[0017] The first bus travel time prediction module is used to predict the bus travel time of the queried bus route using the bus route map, the bus stop map, the queried bus route, the stop popularity information, and the current time information.

[0018] According to another aspect of this disclosure, a training apparatus for a travel time prediction model is provided, the apparatus comprising:

[0019] The second data determination module is used to construct bus route maps and bus stop maps based on road network data, public transport network data, and traffic condition data.

[0020] The second data acquisition module is used to acquire the query bus route between the origin and the destination, the station popularity information of the stations in the query bus route, and the current time information;

[0021] The second bus travel time prediction module is used to predict the bus travel time of the queried bus route using the bus route map, the bus stop map, the queried bus route, the stop popularity information, and the current time information.

[0022] The first model training module is used to train a travel time prediction model based on the bus travel time and the first label data of the queried bus route.

[0023] According to another aspect of this disclosure, an electronic device is provided, the electronic device comprising:

[0024] At least one processor; and

[0025] A memory communicatively connected to the at least one processor; wherein,

[0026] The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the driving time prediction method or the driving time prediction model training method according to any embodiment of this disclosure.

[0027] According to another aspect of this disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause a computer to execute the driving time prediction method or the driving time prediction model training method described in any embodiment of this disclosure.

[0028] According to another aspect of this disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the driving time prediction method or the driving time prediction model training method according to any embodiment of this disclosure.

[0029] The technology disclosed herein can improve the accuracy of bus travel time prediction.

[0030] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0031] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein:

[0032] Figure 1 This is a flowchart of a travel time prediction method provided according to an embodiment of the present disclosure;

[0033] Figure 2 This is a flowchart of yet another travel time prediction method provided according to an embodiment of the present disclosure;

[0034] Figure 3 This is a flowchart of a training method for a travel time prediction model provided according to an embodiment of this disclosure;

[0035] Figure 4 This is a schematic diagram illustrating the training process of a travel time prediction model provided according to an embodiment of the present disclosure;

[0036] Figure 5 This is a schematic diagram of a travel time prediction device provided according to an embodiment of the present disclosure;

[0037] Figure 6 This is a schematic diagram of the structure of a training device for a travel time prediction model provided according to an embodiment of the present disclosure;

[0038] Figure 7 This is a block diagram of an electronic device used to implement the driving time prediction method or the driving time prediction model training method of the embodiments of this disclosure. Detailed Implementation

[0039] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0040] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention 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 the invention 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 a 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.

[0041] Furthermore, it should be noted that the collection, storage, use, processing, transmission, provision, and disclosure of road network data, public transport network data, traffic condition data, bus route queries, and station popularity information involved in the technical solution of this invention all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0042] Figure 1 This is a flowchart illustrating a travel time prediction method according to an embodiment of this disclosure. This method is applicable to scenarios such as bus route planning, real-time bus updates, and electronic bus stop displays, where the goal is to quickly and accurately determine bus travel times. The method can be executed by a travel time prediction device, which can be implemented in software and / or hardware and integrated into an electronic device that carries the travel time prediction function, such as a server. Figure 1 As shown, the travel time prediction method in this embodiment may include:

[0043] S101 constructs bus route maps and bus stop maps based on road network data, public transport network data, and traffic condition data.

[0044] In this embodiment, road network data refers to road data of a city or region, including but not limited to national highways, provincial highways, expressways, municipal administrative roads, urban expressways, first-class roads, township and village roads, and other roads. Public transport network data refers to public transport route data and station data of a city or region. Traffic condition data refers to road condition data of urban roads within each time slice, such as congestion and traffic accidents; it should be noted that a time slice refers to a period of time, such as 5 minutes or half an hour.

[0045] A bus route map is a graph structure extracted from road network data and bus network data, showing the relationships between bus routes and stops. Nodes represent stops, and edges between nodes represent routes between stops; it reflects the correlation between the road network and the bus network. A bus stop map is a graph structure extracted from road network data, bus network data, and traffic condition data, showing the relationships between different stops.

[0046] An alternative approach is to use graph neural networks to transform road network data, public transport network data, and traffic condition data into graph structures to obtain public transport route maps and public transport stop maps.

[0047] S102, obtain the bus route query between the origin and destination, the station popularity information of the bus route, and the current time information.

[0048] In this embodiment, querying a bus route refers to the sequence of stops for a bus route between the origin and the destination, including the stops that need to be passed through; where a stop refers to a bus stop in the bus route.

[0049] Station popularity information reflects the popularity of bus stops and can include passenger flow and / or vehicle density. It can be understood that station popularity information can indirectly reflect the impact of passenger flow and vehicle density on bus travel time, thus laying the foundation for subsequent bus travel time predictions.

[0050] The so-called current time information includes the current time and the time characteristics to which the current time belongs, including weekdays, rest days, peak hours, etc.

[0051] Specifically, it can obtain the starting point and destination entered by the user on the mobile terminal, determine the corresponding bus routes between the starting point and destination from the database, use the queried bus routes as the query bus routes, and then obtain the station popularity information and current time information of the stations in the query bus routes.

[0052] S103 uses bus route maps, bus stop maps, bus route queries, stop popularity information, and current time information to predict the bus travel time for the queried bus route.

[0053] In this embodiment, bus travel time refers to the travel time of the bus in the bus route query.

[0054] One alternative approach involves vector processing of the bus route map, bus stop map, queried bus route, stop popularity information, and current time information to obtain corresponding vectors. Then, based on a travel time prediction network, regression prediction is performed on each vector to obtain the bus travel time for the queried bus route. The travel time prediction network can be a fully connected network.

[0055] The technical solution provided in this disclosure constructs a bus route map and a bus stop map based on road network data, public transport network data, and traffic condition data. It then obtains information such as the query bus route between the origin and destination, the popularity information of stops along the query bus route, and the current time. Finally, it uses the bus route map, bus stop map, query bus route, stop popularity information, and current time information to predict the bus travel time for the query bus route. This technical solution improves the accuracy of bus travel time prediction by combining static data such as road network data and public transport network data with dynamic data such as traffic condition data, query bus routes, stop popularity information, and current time information.

[0056] Based on the above embodiments, as an optional approach of this disclosure, a bus route map and a bus stop map are constructed according to road network data, public transport network data, and traffic condition data, including: constructing a bus route map according to road network data and public transport network data; and constructing a bus stop map according to road network data, public transport network data, and traffic condition data.

[0057] Specifically, based on a preset data conversion method, road network data and public transport network data can be transformed into a graph structure to obtain a public transport route map. Then, based on the preset data conversion method, road network data, public transport network data, and traffic condition data can be transformed into a graph structure to construct a public transport stop map. It should be noted that this embodiment does not specifically limit the preset data conversion method.

[0058] Understandably, using road network data, public transport network data, and traffic condition data to determine public transport route maps and public transport stop maps can yield graph structure data, which facilitates the subsequent characterization of relationships between routes and between stops.

[0059] Figure 2This is a flowchart of another travel time prediction method provided according to an embodiment of this disclosure. Based on the above embodiments, this embodiment further optimizes the process of "predicting the travel time of a queried bus route using a bus route map, bus stop map, querying bus routes, stop popularity information, and current time information," providing an optional implementation scheme. For example... Figure 2 As shown, the travel time prediction method in this embodiment may include:

[0060] S201 constructs bus route maps and bus stop maps based on road network data, public transport network data, and traffic condition data.

[0061] S202, retrieve the bus route query between the origin and destination, the station popularity information of the bus route, and the current time information.

[0062] S203, based on graph attention network, extracts features from bus route map to obtain bus route representation.

[0063] In this embodiment, the bus route representation is used to characterize the correlation between bus routes; it can be represented in vector or matrix form.

[0064] Specifically, the bus route map can be input into a Graph Attention Network (GAT) and processed to obtain a bus route representation.

[0065] S204, based on a spatiotemporal attention network, extracts features from a bus stop map to obtain three-dimensional spatiotemporal features between different stops.

[0066] In this embodiment, three-dimensional spatiotemporal features are used to characterize the correlation between different stations in different time slices, and can be represented in matrix or vector form.

[0067] Specifically, bus stop maps from different time slices can be input into a spatiotemporal attention network. The network learns the spatiotemporal relationships between different stops to obtain the three-dimensional spatiotemporal features between different stops.

[0068] S205 extracts features from the query bus routes, station popularity information, and current time information to obtain the dynamic features of the query bus routes.

[0069] In this embodiment, the route dynamic features are used to query the dynamic characteristics of bus routes in terms of time and space, and can be in matrix or vector form.

[0070] One alternative approach is to vectorize the queried bus route, station popularity information, and current time information separately to obtain queried bus route vectors, station popularity vectors, and current time vectors. Then, based on a route feature extraction network, features can be extracted from the queried bus route vector, station popularity vector, and current time vector respectively to obtain queried bus route features, station popularity features, and current time features. Furthermore, based on a preset feature fusion method, the queried bus route features, station popularity features, and current time features can be fused to obtain dynamic route features. For example, the queried bus route features, station popularity features, and current time features can be concatenated or added together to obtain dynamic route features.

[0071] S206 predicts the bus travel time for the queried bus route based on the bus route representation, three-dimensional spatiotemporal characteristics, and route dynamic characteristics.

[0072] Specifically, the bus route representation, three-dimensional spatiotemporal features, and route dynamic features can be input into the travel time prediction network to predict the bus travel time for the queried bus route.

[0073] The technical solution provided in this disclosure constructs a bus route map and a bus stop map based on road network data, public transport network data, and traffic condition data. It then obtains the bus routes between origin and destination, the station popularity information of the bus routes, and the current time information. Next, based on a graph attention network, it extracts features from the bus route map to obtain a bus route representation. Based on a spatiotemporal attention network, it extracts features from the bus stop map to obtain three-dimensional spatiotemporal features between different stations. Simultaneously, it extracts features from the bus routes, station popularity information, and current time information to obtain the dynamic features of the bus routes. Finally, based on the bus route representation, the three-dimensional spatiotemporal features, and the dynamic features, it predicts the bus travel time for the query bus route. This technical solution introduces a graph attention network for feature extraction from the bus route map, which not only captures the fine-grained characteristics of bus routes but also fully explores the correlations between different routes. Simultaneously, the introduction of a time-based attention network fully utilizes the joint spatiotemporal relationships between different stations in the public transport network, thereby improving the accuracy of bus travel time prediction based on multi-dimensional features.

[0074] Based on the above embodiments, as an optional approach of this disclosure, predicting the bus travel time of a queried bus route according to the bus route representation, three-dimensional spatiotemporal features, and route dynamic features includes: fusing the bus route representation and route dynamic features to obtain fused features; capturing the temporal dynamic features between the three-dimensional spatiotemporal features and the station's three-dimensional spatiotemporal features based on a three-dimensional spatiotemporal attention network, according to the three-dimensional spatiotemporal features and the fused features; and predicting the bus travel time of the queried bus route using the bus route representation, route dynamic features, and temporal dynamic features.

[0075] Among them, fusion features refer to the features obtained by further fusing bus route representation and route dynamic features, which are used to enrich the features of bus route queries; they can be represented in matrix or vector form. Time dynamic features are used to query the correlation between bus routes and the time characteristics of the stops they pass through, and can also be represented in matrix or vector form.

[0076] Specifically, based on a preset feature fusion method, bus route representations and dynamic route features can be fused. For example, the bus route representation and dynamic route features can be concatenated or added together, and the result is used as the fused feature. Then, the fused feature can be used as the input query, and the 3D spatiotemporal features of the stations can be used as the key / value pair. The fused feature and the 3D spatiotemporal features are then input into a 3D spatiotemporal attention network to obtain the temporal dynamic features between the queried bus route and the 3D spatiotemporal features of the stations it passes through. Furthermore, based on a travel time prediction network, the bus route representation, dynamic route features, and temporal dynamic features can be used to predict the bus travel time of the queried bus route.

[0077] Understandably, introducing a three-dimensional spatiotemporal attention network to capture the temporal dynamics between the query bus route and the three-dimensional spatiotemporal features of the stops it passes through can fully explore the dynamic correlation between the temporal features of the query bus route and the stops it passes through, thus laying the foundation for predicting travel time.

[0078] Based on the above embodiments, as an optional method of this disclosure, feature extraction is performed on the query bus route, station popularity information, and current time information to obtain the dynamic features of the query bus route, including: fusing the query bus route, station popularity information, and current time information to obtain fused data; and extracting features from the fused data to obtain the dynamic features of the query bus route.

[0079] Among them, fused data refers to data obtained by combining information on bus routes, station popularity, and current time.

[0080] Specifically, based on certain rules, information such as bus route queries, station popularity, and current time can be fused. For example, these data can be concatenated and used as the fused data. Next, the fused data can be vectorized to obtain a fused vector. Then, based on a route feature extraction network, features can be extracted from the fused data to obtain the dynamic features of the queried bus route. The feature extraction network can be based on a convolutional neural network, such as a fully connected network.

[0081] Understandably, by integrating information on bus routes, station popularity, and current time, the real-time dynamic characteristics of bus route queries can be represented, providing data support for determining the travel time of bus routes.

[0082] Based on the above embodiments, as an optional approach of this disclosure, after extracting features from the bus stop map based on a spatiotemporal attention network to obtain the three-dimensional spatiotemporal features between different stops, the method further includes: determining the driving route between the starting point and the destination; and using the three-dimensional spatiotemporal features to predict the vehicle travel time of the driving route.

[0083] Driving routes refer to the driving routes between the starting point and the destination. It should be noted that driving routes are different from public transportation routes; generally, public transportation routes involve more road segments than driving routes.

[0084] Specifically, the system obtains the driving route between the starting point and the destination within the current time period from the front end, extracts features from the driving route, and then uses a driving time prediction network to predict the driving time of the route using the driving route features and 3D spatiotemporal features. Furthermore, it can also obtain the driving time of the route within a future time period.

[0085] Understandably, this invention can predict not only the travel time of buses for public transport routes but also the travel time of vehicles for driving routes at different times, thus enabling multi-task prediction.

[0086] Figure 3 This is a flowchart illustrating a training method for a travel time prediction model according to an embodiment of this disclosure. This embodiment is applicable to scenarios such as bus route planning, real-time bus updates, and electronic bus stop displays, where the goal is to quickly and accurately determine bus travel times. This method can be executed by a training device for the travel time prediction model, which can be implemented in software and / or hardware and integrated into an electronic device, such as a server, that carries the training function of the travel time prediction model. Figure 3As shown, the training method for the travel time prediction model in this embodiment may include:

[0087] S301 constructs bus route maps and bus stop maps based on road network data, public transport network data, and traffic condition data.

[0088] In this embodiment, road network data refers to road data of a city or region, including but not limited to national highways, provincial highways, expressways, municipal administrative roads, urban expressways, first-class roads, township and village roads, and other roads. Public transport network data refers to public transport route data and station data of a city or region. Traffic condition data refers to road condition data of urban roads within each time slice, such as congestion and traffic accidents; it should be noted that a time slice refers to a period of time, such as 5 minutes or half an hour.

[0089] A bus route map is a graph structure extracted from road network data and bus network data, showing the relationships between bus routes and stops. Nodes represent stops, and edges between nodes represent routes between stops; it reflects the correlation between the road network and the bus network. A bus stop map is a graph structure extracted from road network data, bus network data, and traffic condition data, showing the relationships between different stops.

[0090] An alternative approach is to use graph neural networks to transform road network data, public transport network data, and traffic condition data into graph structures to obtain public transport route maps and public transport stop maps.

[0091] Another option is to construct a bus route map based on road network data and public transport network data; and to construct a bus stop map based on road network data, public transport network data, and traffic condition data.

[0092] S302, retrieve the bus route query between the origin and destination, the station popularity information of the bus route, and the current time information.

[0093] In this embodiment, querying a bus route refers to the sequence of stops for a bus route between the origin and the destination, including the stops that need to be passed through; where a stop refers to a bus stop in the bus route.

[0094] Station popularity information reflects the popularity of bus stops and can include passenger flow and / or vehicle density. It can be understood that station popularity information can indirectly reflect the impact of passenger flow and vehicle density on bus travel time, thus laying the foundation for subsequent bus travel time predictions.

[0095] The so-called current time information includes the current time and the time characteristics to which the current time belongs, including weekdays, rest days, peak hours, etc.

[0096] Specifically, it can obtain the starting point and destination entered by the user on the mobile terminal, determine the corresponding bus routes between the starting point and destination from the database, use the queried bus routes as the query bus routes, and then obtain the station popularity information and current time information of the stations in the query bus routes.

[0097] S303 uses bus route maps, bus stop maps, bus route queries, stop popularity information, and current time information to predict the bus travel time for the queried bus route.

[0098] In this embodiment, bus travel time refers to the travel time of the bus in the bus route query.

[0099] One alternative approach involves vector processing of the bus route map, bus stop map, queried bus route, stop popularity information, and current time information to obtain corresponding vectors. Then, based on a travel time prediction network, regression prediction is performed on each vector to obtain the bus travel time for the queried bus route. The travel time prediction network can be a fully connected network.

[0100] S304. Based on the bus travel time and the first label data of the bus route query, a travel time prediction model is trained.

[0101] In this embodiment, the first label data refers to the actual bus travel time for the queried bus route. The so-called travel time prediction model is used to predict travel time, which can be used for predicting bus travel time for bus query routes, and also for predicting vehicle travel time for driving routes; the optional travel time prediction model may include graph attention network, spatiotemporal graph attention network and 3D spatiotemporal attention network, route feature extraction network and travel time prediction network.

[0102] Specifically, based on a first preset loss function, a first training loss can be determined using bus travel time and first label data of queried bus routes. The travel time prediction model is then iteratively trained using this first training loss until a first training stopping condition is met, at which point training of the travel time prediction model ceases. The first training loss is determined based on the bus travel time and the first label data. The first training stopping condition may include the number of iterations meeting a first set number, or the first training loss stabilizing within a first set range; the first set number and the first set range can be set by those skilled in the art according to actual needs. It should be noted that this embodiment does not specifically limit the first preset loss function; for example, it could be a cross-entropy loss function.

[0103] The technical solution provided in this disclosure constructs a bus route map and a bus stop map based on road network data, public transport network data, and traffic condition data. It then obtains the bus route between the origin and destination, the station popularity information of the bus routes, and the current time information. Using the bus route map, bus stop map, bus route data, station popularity information, and current time information, it predicts the bus travel time for the queried bus route. Finally, it trains a travel time prediction model based on the bus travel time and the first label data of the queried bus route. This technical solution uses static data such as road network data and public transport network data, as well as dynamic data such as traffic condition data, queried bus routes, station popularity information, and current time information. By training a travel time prediction model based on dynamic and static data from different dimensions, it improves the accuracy of bus travel time prediction.

[0104] Based on the above embodiments, as an optional method of this disclosure, a bus route map, a bus stop map, query bus routes, stop popularity information, and current time information are used to predict the bus travel time of a query bus route. This includes: extracting features from the bus route map based on a graph attention network to obtain a bus route representation; extracting features from the bus stop map based on a spatiotemporal attention network to obtain three-dimensional spatiotemporal features between different stops; extracting features from the query bus route, stop popularity information, and current time information to obtain the dynamic features of the query bus route; and predicting the bus travel time of the query bus route based on the bus route representation, the three-dimensional spatiotemporal features, and the dynamic features.

[0105] Among them, bus route representation is used to characterize the correlation between bus routes; it can be represented in vector or matrix form. Three-dimensional spatiotemporal features are used to characterize the correlation between different stations within different time slices; they can also be represented in matrix or vector form. Route dynamic features are used to query the dynamic characteristics of bus routes, such as time and space; they can also be represented in matrix or vector form.

[0106] Specifically, the bus route map can be input into a Graph Attention Network (GAT) for processing to obtain a bus route representation. Then, bus stop maps from different time slices can be input into a spatiotemporal attention network. The network learns the spatiotemporal relationships between different stops, obtaining three-dimensional spatiotemporal features between them. Next, the queried bus route, stop popularity information, and current time information can be vectorized to obtain queried bus route vectors, stop popularity vectors, and current time vectors. Then, based on a route feature extraction network, features can be extracted from the queried bus route vector, stop popularity vector, and current time vector to obtain queried bus route features, stop popularity features, and current time features. Finally, based on a preset feature fusion method, the queried bus route features, stop popularity features, and current time features can be fused to obtain dynamic route features. For example, the queried bus route features, stop popularity features, and current time features can be concatenated or added together to obtain dynamic route features.

[0107] Understandably, introducing graph attention networks to extract features from bus route maps can not only capture the fine-grained characteristics of bus routes, but also fully explore the correlations between different routes. At the same time, introducing time-attention networks can make full use of the joint spatiotemporal relationships between different stations in the bus network, thereby improving the accuracy of bus travel time prediction based on multi-dimensional features.

[0108] Based on the above embodiments, as an optional approach of this disclosure, after extracting features from the bus stop map based on a spatiotemporal attention network to obtain the three-dimensional spatiotemporal features between different stops, the method further includes: determining the driving route between the starting point and the destination; using the three-dimensional spatiotemporal features to predict the vehicle travel time of the driving route; and training the travel time prediction model based on the second label data of the vehicle travel time and the driving route.

[0109] The driving route refers to the driving path between the origin and destination. It's important to note that driving routes differ from public transport routes; generally, public transport routes involve more road segments than driving routes. The second tag data refers to the actual vehicle travel time for the driving route.

[0110] Specifically, the driving route between the starting point and the destination within the current time period is obtained from the front end. Features are extracted from the driving route to obtain driving route features. Then, based on the driving time prediction network, the driving time of the driving route is predicted using the driving route features and 3D spatiotemporal features to obtain the vehicle travel time for the driving route. Furthermore, the vehicle travel time of the driving route within a future time period can also be obtained. Next, based on a second preset loss function, a second training loss is determined using the second label data of the vehicle travel time and the driving route. The driving time prediction model is then iteratively trained using the second training loss until a second training stopping condition is met.

[0111] The second training loss is determined based on vehicle travel time and second label data. The second training stopping condition may include the number of iterations meeting a second preset number, or the second training loss stabilizing within a second preset range; the second preset number and the second preset range can be set by those skilled in the art according to actual needs. It should be noted that this embodiment does not specifically limit the second preset loss function; for example, it could be the cross-entropy loss function.

[0112] Understandably, this disclosure allows for multi-task training of the travel time prediction model. The main task is to predict the travel time of public transportation routes, while the auxiliary task is to predict the vehicle travel time of driving routes at different times, thus enabling multi-task prediction.

[0113] Figure 4 This is a schematic diagram illustrating the training process of a travel time prediction model provided according to an embodiment of this disclosure. Based on the above embodiments, this embodiment provides an example of the training process for a travel time prediction model. Optionally, the travel time prediction model includes a graph attention network, a spatiotemporal attention network, a 3D spatiotemporal attention network, a route feature extraction network, and a travel time prediction network; wherein, the graph attention network is used to extract features from the bus route map; the spatiotemporal attention network is used to extract features from the bus stop map; the 3D spatiotemporal attention network is used to capture the temporal dynamic features between the 3D spatiotemporal features of the queried bus route and stops; the route feature extraction network is used to extract route dynamic features from the queried bus route, stop popularity information, and current time information; and the travel time prediction network is used to perform multi-task prediction of bus travel time and vehicle travel time.

[0114] Specifically, the bus route map can be input into a graph attention network for processing to obtain the bus route representation. The bus stop map can be input into a spatiotemporal attention network to obtain the three-dimensional spatiotemporal features between different stops. The queried bus route, stop popularity information, and current time information can be input into a route feature extraction network for feature extraction to obtain the dynamic features of the queried bus route. Then, the bus route representation and dynamic features are fused to obtain fused features. Based on the three-dimensional spatiotemporal attention network, weighted attention operations are performed on the fused features and the three-dimensional spatiotemporal features to obtain the temporal dynamic features between the queried bus route and the three-dimensional spatiotemporal features of the stops. Furthermore, the bus route representation, route dynamic features, and temporal dynamic features are integrated to obtain integrated features. Finally, the integrated features are input into a travel time prediction network to obtain the bus travel time for the queried bus route. Using the bus travel time and the first label data of the queried bus route, a first training loss is determined, and the travel time prediction model is trained using this first training loss.

[0115] Furthermore, three-dimensional spatiotemporal features and driving route features can be input into the driving time prediction network to obtain the vehicle driving time of the driving route. Then, the second label data of vehicle driving time and driving route are used to determine the second training loss, and the driving time prediction model is trained using the second training loss.

[0116] Figure 5 This is a schematic diagram of a travel time prediction device according to an embodiment of this disclosure. This embodiment is applicable to scenarios such as bus route planning, real-time bus updates, and electronic bus stop displays, where the goal is to quickly and accurately determine bus travel times. The device can be implemented using software and / or hardware and can be integrated into electronic devices that perform travel time prediction functions, such as servers.

[0117] like Figure 5 As shown, the travel time prediction device 500 in this embodiment may include:

[0118] The first diagram structure determination module 501 is used to construct bus route maps and bus stop maps based on road network data, public transport network data, and traffic condition data.

[0119] The first data acquisition module 502 is used to acquire the query bus route between the origin and the destination, the station popularity information of the bus route, and the current time information;

[0120] The first bus travel time prediction module 503 is used to predict the bus travel time of the queried bus route by using bus route map, bus stop map, query bus route, stop popularity information, and current time information.

[0121] The technical solution provided in this disclosure constructs a bus route map and a bus stop map based on road network data, public transport network data, and traffic condition data. It then obtains information such as the query bus route between the origin and destination, the popularity information of stops along the query bus route, and the current time. Finally, it uses the bus route map, bus stop map, query bus route, stop popularity information, and current time information to predict the bus travel time for the query bus route. This technical solution improves the accuracy of bus travel time prediction by combining static data such as road network data and public transport network data with dynamic data such as traffic condition data, query bus routes, stop popularity information, and current time information.

[0122] Furthermore, the first diagram structure determination module 501 is specifically used for:

[0123] Construct a bus route map based on road network data and public transport network data;

[0124] A bus stop map is constructed based on road network data, public transport network data, and traffic condition data.

[0125] Furthermore, the first bus travel time prediction module 503 includes:

[0126] The first bus route representation determination unit is used to extract features from the bus route map based on a graph attention network to obtain the bus route representation;

[0127] The first three-dimensional spatiotemporal feature determination unit is used to extract features from the bus stop map based on the spatiotemporal attention network to obtain the three-dimensional spatiotemporal features between different stops;

[0128] The first route dynamic feature determination unit is used to extract features from the queried bus route, station popularity information, and current time information to obtain the route dynamic features of the queried bus route.

[0129] The first bus travel time prediction unit is used to predict the bus travel time of the queried bus route based on the bus route representation, three-dimensional spatiotemporal characteristics, and route dynamic characteristics.

[0130] Furthermore, the first bus travel time prediction unit is specifically used for:

[0131] The bus route representation and route dynamic features are fused to obtain the fused features;

[0132] Based on a three-dimensional spatiotemporal attention network, the temporal dynamic features between the three-dimensional spatiotemporal features and the fusion features of the queried bus routes and stations are captured according to the three-dimensional spatiotemporal features and the fusion features.

[0133] By using bus route representation, route dynamic features, and time dynamic features, the system predicts the bus travel time for queried bus routes.

[0134] Furthermore, the first line dynamic characteristic determination unit is specifically used for:

[0135] The system integrates information on bus routes, popular bus stops, and current time to obtain fused data.

[0136] Feature extraction is performed on the fused data to obtain the dynamic features of the queried bus routes.

[0137] Furthermore, the device also includes a vehicle travel time prediction module for:

[0138] After extracting features from the bus stop map based on a spatiotemporal attention network to obtain the three-dimensional spatiotemporal features between different stops, the driving route between the starting point and the destination is determined.

[0139] Using three-dimensional spatiotemporal features, the vehicle travel time for driving routes is predicted.

[0140] Furthermore, site popularity information includes site pedestrian traffic and / or vehicle density.

[0141] Figure 6 This is a schematic diagram of a training device for a travel time prediction model provided according to an embodiment of this disclosure. This embodiment is applicable to scenarios such as bus route planning, real-time bus updates, and electronic bus stop displays, where the goal is to quickly and accurately determine bus travel times. The device can be implemented using software and / or hardware and can be integrated into an electronic device, such as a server, that carries the training function of the travel time prediction model. Figure 6 As shown, the training device 600 for the travel time prediction model in this embodiment may include:

[0142] The second data determination module 601 is used to construct a bus route map and a bus stop map based on road network data, public transport network data and traffic condition data.

[0143] The second data acquisition module 602 is used to acquire the query bus route between the origin and the destination, the station popularity information of the bus route, and the current time information;

[0144] The second bus travel time prediction module 603 is used to predict the bus travel time of the queried bus route by using bus route map, bus stop map, query bus route, stop popularity information, and current time information.

[0145] The first model training module 604 is used to train a travel time prediction model based on the first label data of bus travel time and query bus routes.

[0146] The technical solution provided in this disclosure constructs a bus route map and a bus stop map based on road network data, public transport network data, and traffic condition data. It then obtains the bus route between the origin and destination, the station popularity information of the bus routes, and the current time information. Using the bus route map, bus stop map, bus route data, station popularity information, and current time information, it predicts the bus travel time for the queried bus route. Finally, it trains a travel time prediction model based on the bus travel time and the first label data of the queried bus route. This technical solution uses static data such as road network data and public transport network data, as well as dynamic data such as traffic condition data, queried bus routes, station popularity information, and current time information. By training a travel time prediction model based on dynamic and static data from different dimensions, it improves the accuracy of bus travel time prediction.

[0147] Furthermore, the second bus travel time prediction module 603 includes:

[0148] The second bus route representation determination unit is used to extract features from the bus route map based on a graph attention network to obtain the bus route representation.

[0149] The second three-dimensional spatiotemporal feature determination unit is used to extract features from the bus stop map based on the spatiotemporal attention network to obtain the three-dimensional spatiotemporal features between different stops;

[0150] The second route dynamic feature determination unit is used to extract features from the queried bus routes, station popularity information, and current time information to obtain the route dynamic features of the queried bus routes.

[0151] The second bus travel time prediction unit is used to predict the bus travel time of the queried bus route based on the bus route representation, three-dimensional spatiotemporal characteristics, and route dynamic characteristics.

[0152] Furthermore, the device also includes a second model training module for:

[0153] Based on a spatiotemporal attention network, features are extracted from the bus stop map to obtain the three-dimensional spatiotemporal features between different stops, and then the driving route between the starting point and the destination is determined.

[0154] Using three-dimensional spatiotemporal features, predict vehicle travel time for driving routes;

[0155] A travel time prediction model is trained based on second-label data of vehicle travel time and driving route.

[0156] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0157] Figure 7 This is a block diagram of an electronic device used to implement the driving time prediction method or driving time prediction model training method according to embodiments of the present disclosure. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0158] like Figure 7 As shown, the electronic device 700 includes a computing unit 701, which can perform various appropriate actions and processes based on a computer program stored in a read-only memory (ROM) 702 or a computer program loaded from a storage unit 708 into a random access memory (RAM) 703. The RAM 703 may also store various programs and data required for the operation of the electronic device 700. The computing unit 701, ROM 702, and RAM 703 are interconnected via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0159] Multiple components in electronic device 700 are connected to I / O interface 705, including: input unit 706, such as keyboard, mouse, etc.; output unit 707, such as various types of displays, speakers, etc.; storage unit 708, such as disk, optical disk, etc.; and communication unit 709, such as network card, modem, wireless transceiver, etc. Communication unit 709 allows electronic device 700 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0160] The computing unit 701 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 701 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 701 performs the various methods and processes described above, such as driving time prediction methods or driving time prediction model training methods. For example, in some embodiments, the driving time prediction method or driving time prediction model training method can be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 708. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 700 via ROM 702 and / or communication unit 709. When the computer program is loaded into RAM 703 and executed by the computing unit 701, one or more steps of the driving time prediction method or driving time prediction model training method described above can be performed. Alternatively, in other embodiments, the computing unit 701 may be configured by any other suitable means (e.g., by means of firmware) to perform a driving time prediction method or a driving time prediction model training method.

[0161] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0162] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0163] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0164] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0165] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0166] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0167] Artificial intelligence (AI) is the study of enabling computers to simulate certain human thought processes and intelligent behaviors (such as learning, reasoning, thinking, and planning). It encompasses both hardware and software technologies. AI hardware technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, and big data processing. AI software technologies mainly include computer vision, speech recognition, natural language processing, machine learning / deep learning, big data processing, and knowledge graph technologies.

[0168] Cloud computing refers to a technology system that enables access to a shared pool of physical or virtual resources via a network. These resources can include servers, operating systems, networks, software, applications, and storage devices, and can be deployed and managed on demand and in a self-service manner. Cloud computing technology can provide efficient and powerful data processing capabilities for applications such as artificial intelligence and blockchain, as well as for model training.

[0169] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0170] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for predicting travel time, comprising: Based on road network data, public transport network data, and traffic condition data, construct a public transport route map and a public transport stop map; Obtain the bus route between the origin and destination, the station popularity information of the bus route, and the current time information; Based on a graph attention network, feature extraction is performed on the bus route map to obtain a bus route representation. Based on a spatiotemporal attention network, feature extraction is performed on the bus stop map to obtain the three-dimensional spatiotemporal features between different stops. Feature extraction is performed on the queried bus route, the station popularity information, and the current time information to obtain the dynamic features of the queried bus route; Based on the bus route representation, the three-dimensional spatiotemporal features, and the route dynamic features, predict the bus travel time for the queried bus route; The method of extracting features from the bus stop map based on a spatiotemporal attention network to obtain the three-dimensional spatiotemporal features between different stops further includes: Determine the driving route between the starting point and the destination; Using the aforementioned three-dimensional spatiotemporal features, the vehicle travel time for the driving route is predicted.

2. The method according to claim 1, wherein, The process of constructing bus route maps and bus stop maps based on road network data, public transport network data, and traffic condition data includes: Construct a bus route map based on road network data and public transport network data; A bus stop map is constructed based on the road network data, the public transport network data, and the traffic condition data.

3. The method according to claim 1, wherein, The step of predicting the bus travel time of the queried bus route based on the bus route representation, the three-dimensional spatiotemporal features, and the route dynamic features includes: The bus route representation and the route dynamic features are fused to obtain fused features; Based on a three-dimensional spatiotemporal attention network, the temporal dynamic features between the three-dimensional spatiotemporal features and the fused features are captured. Using the bus route representation, the route dynamic features, and the time dynamic features, the bus travel time of the queried bus route is predicted.

4. The method according to claim 1, wherein, The step of extracting features from the queried bus route, the station popularity information, and the current time information to obtain the dynamic features of the queried bus route includes: The queried bus route, the station popularity information, and the current time information are fused to obtain fused data; Feature extraction is performed on the fused data to obtain the dynamic features of the queried bus route.

5. The method according to any one of claims 1-4, wherein, The site popularity information includes the site's pedestrian traffic and / or vehicle density.

6. A training method for a travel time prediction model, comprising: Based on road network data, public transport network data, and traffic condition data, construct a public transport route map and a public transport stop map; Obtain the bus route between the origin and destination, the station popularity information of the bus route, and the current time information; Based on a graph attention network, feature extraction is performed on the bus route map to obtain a bus route representation. Based on a spatiotemporal attention network, feature extraction is performed on the bus stop map to obtain the three-dimensional spatiotemporal features between different stops. Feature extraction is performed on the queried bus route, the station popularity information, and the current time information to obtain the dynamic features of the queried bus route; Based on the bus route representation, the three-dimensional spatiotemporal features, and the route dynamic features, predict the bus travel time for the queried bus route; Based on the bus travel time and the first label data of the queried bus route, train a travel time prediction model; The method of extracting features from the bus stop map based on a spatiotemporal attention network to obtain the three-dimensional spatiotemporal features between different stops further includes: Determine the driving route between the starting point and the destination; Using the aforementioned three-dimensional spatiotemporal features, the vehicle travel time for the driving route is predicted; A driving time prediction model is trained based on the vehicle's travel time and the second label data of the driving route.

7. A travel time prediction device, comprising: The first diagram structure determination module is used to construct bus route maps and bus stop maps based on road network data, public transport network data, and traffic condition data. The first data acquisition module is used to acquire the query bus route between the origin and the destination, the station popularity information of the stations in the query bus route, and the current time information; The first bus travel time prediction module includes: The first bus route representation determination unit is used to extract features from the bus route map based on a graph attention network to obtain a bus route representation. The first three-dimensional spatiotemporal feature determination unit is used to extract features from the bus stop map based on a spatiotemporal attention network to obtain the three-dimensional spatiotemporal features between different stops. The first route dynamic feature determination unit is used to extract features from the queried bus route, the station popularity information, and the current time information to obtain the route dynamic features of the queried bus route. The first bus travel time prediction unit is used to predict the bus travel time of the queried bus route based on the bus route representation, the three-dimensional spatiotemporal features and the route dynamic features. The device also includes a vehicle travel time prediction module, used for: After extracting features from the bus stop map based on a spatiotemporal attention network to obtain the three-dimensional spatiotemporal features between different stops, the driving route between the starting point and the destination is determined. Using the aforementioned three-dimensional spatiotemporal features, the vehicle travel time for the driving route is predicted.

8. The apparatus according to claim 7, wherein, The first diagram structure determination module is specifically used for: Construct a bus route map based on road network data and public transport network data; A bus stop map is constructed based on the road network data, the public transport network data, and the traffic condition data.

9. The apparatus according to claim 7, wherein, The first bus travel time prediction unit is specifically used for: The bus route representation and the route dynamic features are fused to obtain fused features; Based on a three-dimensional spatiotemporal attention network, the temporal dynamic features between the three-dimensional spatiotemporal features and the fused features are captured. Using the bus route representation, the route dynamic features, and the time dynamic features, the bus travel time of the queried bus route is predicted.

10. The apparatus according to claim 7, wherein, The first line dynamic characteristic determination unit is specifically used for: The queried bus route, the station popularity information, and the current time information are fused to obtain fused data; Feature extraction is performed on the fused data to obtain the dynamic features of the queried bus route.

11. The apparatus according to any one of claims 7-10, wherein, The site popularity information includes the site's pedestrian traffic and / or vehicle density.

12. A training device for a travel time prediction model, comprising: The second data determination module is used to construct bus route maps and bus stop maps based on road network data, public transport network data, and traffic condition data. The second data acquisition module is used to acquire the query bus route between the origin and the destination, the station popularity information of the stations in the query bus route, and the current time information; The second bus travel time prediction module includes: The second bus route representation determination unit is used to extract features from the bus route map based on a graph attention network to obtain a bus route representation. The second three-dimensional spatiotemporal feature determination unit is used to extract features from the bus stop map based on a spatiotemporal attention network to obtain the three-dimensional spatiotemporal features between different stops. The second route dynamic feature determination unit is used to extract features from the queried bus route, the station popularity information, and the current time information to obtain the route dynamic features of the queried bus route. The second bus travel time prediction unit is used to predict the bus travel time of the queried bus route based on the bus route representation, the three-dimensional spatiotemporal features and the route dynamic features. The first model training module is used to train a travel time prediction model based on the bus travel time and the first label data of the queried bus route. The second model training module is used for: Based on a spatiotemporal attention network, feature extraction is performed on the bus stop map to obtain the three-dimensional spatiotemporal features between different stops, and then the driving route between the starting point and the destination is determined. Using the aforementioned three-dimensional spatiotemporal features, the vehicle travel time for the driving route is predicted; A driving time prediction model is trained based on the vehicle's travel time and the second label data of the driving route.

13. An electronic device, comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the driving time prediction method according to any one of claims 1-5, or the training method for the driving time prediction model according to claim 6.

14. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to execute the driving time prediction method according to any one of claims 1-5, or the training method of the driving time prediction model according to claim 6.

15. A computer program product comprising a computer program that, when executed by a processor, implements the driving time prediction method according to any one of claims 1-5, or the training method for the driving time prediction model according to claim 6.

Images