CN115031756A - Travel information determination method and device and computer program product - Google Patents

Abstract

The disclosure provides a method and a device for determining travel information, electronic equipment, a storage medium and a computer program product, relates to the technical field of artificial intelligence, in particular to a deep learning technology, and can be used in a navigation scene. The specific implementation scheme is as follows: determining a starting point and a stopping point in the acquired route calculation request; determining road network time sequence characteristic information of road network data at multiple moments in a time period to which the current moment belongs through a pre-trained characteristic extraction network; determining coding information corresponding to a start-stop point from the road network timing characteristic information; and determining trip information including a connection position between a start point and a stop point and a connection trip mode through a pre-trained multi-task network based on the coding information. The method and the device improve the accuracy of the obtained travel information and the flexibility of the travel mode.

Description

Method, device and computer program product for determining travel information

technical field

The present disclosure relates to the field of artificial intelligence, in particular to deep learning technology, and in particular to a method and device for determining travel information and a training method, device, electronic device, storage medium and computer program product for a travel information determination model, which can be used in navigation scenarios.

Background technique

Obtaining the total travel route through the combination of multiple travel modes can often make the travel process achieve better results in terms of time, money cost, and energy consumption. In the prior art, generally by means of round-robin search, different travel modes are violently combined to obtain different combination schemes, and then the recalled combination schemes are sorted and filtered to obtain higher-quality combined travel schemes. The existing path calculation method through the round-robin search method requires a huge amount of calculation and takes a long time to calculate.

SUMMARY OF THE INVENTION

The present disclosure provides a method and device for determining travel information, and a method, device, electronic device, storage medium, and computer program product for training a travel information determination model.

According to a first aspect, a method for determining travel information is provided, including: determining the starting and ending points in the obtained route calculation request; The road network time series characteristic information of the network data; the coding information corresponding to the start and end points is determined from the road network time series characteristic information; based on the code information, the pre-trained multi-task network is used to determine the connection position and connection travel between the start and end points. mode of travel information.

According to a second aspect, a training method for a travel information determination model is provided, including: acquiring a training sample set, wherein the training samples in the training sample set include trajectory data and connection location labels containing different travel modes, and connection travel modes. Label; determine the road network time series feature information of the road network data at multiple moments in the target time period corresponding to the input trajectory data through the feature extraction network; determine the corresponding start and end points of the input trajectory data from the road network time series feature information Input the coding information into the multi-task network, take the connection position label and connection travel mode label corresponding to the input trajectory data as the expected output of the multi-task network, and train to obtain the travel including the feature extraction network and the multi-task network. Information determines the model.

According to a third aspect, a device for determining travel information is provided, comprising: a first determining unit configured to determine a starting and ending point in the acquired route calculation request; an extraction unit configured to extract a network through a pre-trained feature Determine the road network timing feature information of the road network data at multiple moments in the time period to which the current moment belongs; the coding unit is configured to determine the coding information corresponding to the starting and ending points from the road network timing feature information; the second determining unit, is configured to determine travel information including connection locations between start and end points and connection travel modes through a pre-trained multi-task network based on the encoded information.

According to a fourth aspect, a training device for a travel information determination model is provided, including: an acquisition unit configured to acquire a training sample set, wherein the training samples in the training sample set include trajectory data and connection positions containing different travel modes labels, and labels for connecting travel modes; the training unit is configured to determine, through the feature extraction network, the road network time series feature information of the road network data at multiple moments in the target time period corresponding to the input trajectory data; from the road network time series features The coding information corresponding to the starting and ending points of the input trajectory data is determined in the information; the coding information is input into the multi-task network, and the connection position label and connection travel mode label corresponding to the input trajectory data are used as the expected output of the multi-task network. The travel information determination model including feature extraction network and multi-task network is obtained by training.

According to a fifth aspect, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are processed by the at least one processor The processor executes, so that at least one processor can execute the method described in any implementation manner of the first aspect and the second aspect.

According to a sixth aspect, there is provided a non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the method as described in any implementation manner of the first aspect and the second aspect.

According to a seventh aspect, there is provided a computer program product, comprising: a computer program, when executed by a processor, the computer program implements the method described in any implementation manner of the first aspect and the second aspect.

According to the technology of the present disclosure, a method for determining travel information is provided. For road network time series feature information of road network data at multiple times extracted by a feature extraction network in a travel information determination model, a starting point corresponding to the starting point is determined. Then, through the multi-task network, based on the determined coding information, the connection position between the starting and ending points and the connection travel mode are determined, which improves the accuracy of the obtained travel information and the flexibility of the travel mode.

It should be understood that what is described in this section is not intended to identify key or critical features of embodiments of the disclosure, nor is it intended to limit the scope of the disclosure. Other features of the present disclosure will become readily understood from the following description.

Description of drawings

The accompanying drawings are used for better understanding of the present solution, and do not constitute a limitation to the present disclosure. in:

1 is an exemplary system architecture diagram to which an embodiment of the present disclosure may be applied;

2 is a flowchart of an embodiment of a method for determining travel information according to the present disclosure;

3 is a schematic diagram of an application scenario of the method for determining travel information according to the present embodiment;

4 is a schematic structural diagram of a travel information determination model according to the present disclosure;

FIG. 5 is a flowchart of still another embodiment of a method for determining travel information according to the present disclosure;

6 is a flowchart of an embodiment of a training method for a travel information determination model according to the present disclosure;

7 is a structural diagram of an embodiment of a device for determining travel information according to the present disclosure;

8 is a structural diagram of an embodiment of a training device for a travel information determination model according to the present disclosure;

FIG. 9 is a schematic structural diagram of a computer system suitable for implementing embodiments of the present disclosure.

Detailed ways

Exemplary embodiments of the present disclosure are described below with reference to the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding and should be considered as exemplary only. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions are omitted from the following description for clarity and conciseness.

In the technical solutions of the present disclosure, the collection, storage, use, processing, transmission, provision, and disclosure of the user's personal information involved are all in compliance with relevant laws and regulations, and do not violate public order and good customs.

FIG. 1 shows an exemplary architecture 100 of a method and apparatus for determining travel information, and a method and apparatus for training a travel information determination model to which the present disclosure may be applied.

As shown in FIG. 1 , the system architecture 100 may include terminal devices 101 , 102 , and 103 , a network 104 and a server 105 . The communication connections among the terminal devices 101 , 102 , and 103 constitute a topology network, and the network 104 is used to provide a medium for communication links between the terminal devices 101 , 102 , and 103 and the server 105 . The network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, among others.

The terminal devices 101, 102, and 103 may be hardware devices or software that support network connection for data interaction and data processing. When the terminal devices 101, 102, and 103 are hardware, they can be various electronic devices that support network connection, information acquisition, interaction, display, processing and other functions, including but not limited to smart phones, tablet computers, e-book readers, Laptops and desktops, etc. When the terminal devices 101, 102, and 103 are software, they can be installed in the electronic devices listed above. It can be implemented, for example, as multiple software or software modules for providing distributed services, or as a single software or software module. There is no specific limitation here.

The server 105 may be a server that provides various services, for example, for the route calculation requests sent by the terminal devices 101, 102 and 103, for the route calculation requests of the road network data at multiple times extracted by the feature extraction network in the travel information determination model The network timing feature information is used to determine the accurate coding information corresponding to the starting point, and then the connection position between the starting and ending points and the background processing server for the connection travel mode are determined based on the determined coding information through the multi-task network. For another example, through the training samples provided by the terminal devices 101 , 102 and 103 , a background processing server including a feature extraction network and a travel information determination model of a multi-task network is obtained by training. As an example, the server 105 may be a cloud server.

It should be noted that the server may be hardware or software. When the server is hardware, it can be implemented as a distributed server cluster composed of multiple servers, or can be implemented as a single server. When the server is software, it can be implemented as a plurality of software or software modules (for example, software or software modules for providing distributed services), or can be implemented as a single software or software module. There is no specific limitation here.

It should also be noted that the method for determining travel information and the method for training a model for determining travel information provided by the embodiments of the present disclosure may be executed by a server, a terminal device, or a server and a terminal device in cooperation with each other. Correspondingly, each part (for example, each unit) included in the device for determining travel information and the training device for determining the model for travel information can be all set in the server, or all can be set in the terminal device, and can also be set in the server and the terminal device respectively. middle.

It should be understood that the numbers of terminal devices, networks and servers in FIG. 1 are merely illustrative. There can be any number of terminal devices, networks and servers according to implementation needs. When the electronic device on which the method for determining travel information and the method for training the model for determining travel information does not need to transmit data with other electronic devices, the system architecture may only include the method for determining travel information and the training method for determining the model for travel information. An electronic device (eg a server or terminal device) on which the method runs.

Please refer to FIG. 2, which is a flowchart of a method for determining travel information according to an embodiment of the present disclosure, wherein the process 200 includes the following steps:

Step 201: Determine the start and end points in the obtained route calculation request.

In this embodiment, the execution body of the method for determining travel information (for example, the terminal device or the server in FIG. 1 ) may obtain a route calculation request from a remote location or locally based on a wired network connection method or a wireless network connection method, and determine the route calculation request. The starting and ending points in the obtained route calculation request.

The route calculation request is used to request the calculation of the navigation route between the starting and ending points. As an example, the map navigation application in the above-mentioned execution body provides an application interface for inputting the starting and ending points for the user, and the application interface includes input boxes for the starting and ending points in the starting and ending points, and the user can input the corresponding starting and ending point information in the input boxes, Send a route request.

After receiving the route calculation request, the above-mentioned execution body can analyze the above-mentioned route calculation request, and determine the start and end points included in the route calculation request.

Step 202: Determine, through a pre-trained feature extraction network, road network time series feature information of the road network data at multiple moments within the time period to which the current moment belongs.

In this embodiment, the above-mentioned execution subject may determine, through a pre-trained feature extraction network, road network time series feature information of road network data at multiple times within the time period to which the current moment belongs.

Road network data includes road data under various travel modes, including but not limited to subway network data, bus road network data, driving road network data, walking road network data, and cycling road network data. It can be understood that, in general, the same road data may include multiple attributes. For example, the same road might include transit road attributes and driving road attributes.

Due to factors such as road construction, road control, and tidal lanes, road network data may change at different times. In order to determine the impact of changes in network data on travel information, the above-mentioned executive body can extract the feature information of road network data at multiple moments near the current moment through a pre-trained feature extraction network, and then extract the feature information of road network data at each moment. The timing change information of the road network can be obtained to obtain the timing characteristic information of the road network.

In this embodiment, the pre-trained feature extraction network may be an encoding network with a feature extraction function, including but not limited to a convolutional neural network model, a residual network model, an AlexNet model, a VGG (Visual Geometry Group Network) network) model.

Specifically, the above-mentioned execution body may take days as a period, and divide a day into multiple times (for example, divide a day into 96 times on average). Using the time length corresponding to the preset time period as the time window, multiple times adjacent to the current time are determined. The length of the time period can be set flexibly. As an example, the duration of the time period is 120 minutes.

Step 203: Determine the coding information corresponding to the start and end points from the road network time sequence feature information.

In this embodiment, the above-mentioned execution body may determine the coding information corresponding to the start and end points from the road network time sequence feature information.

The road network time series feature information includes coded information of each position in the road network. In this embodiment, the above-mentioned execution body may determine the coding information of the starting point position and the coding information of the ending point position from the road network time series feature information based on the matching of the position information.

Step 204 , based on the encoded information, determine travel information including the connecting location between the starting and ending points and the connecting travel mode through the pre-trained multi-task network.

In this embodiment, the above-mentioned execution subject may determine travel information including the connecting position between the starting and ending points and the connecting travel mode through a pre-trained multi-task network based on the encoded information.

Specifically, the above-mentioned executive body can input the encoded information of the starting and ending points into the multi-task network, output the connection position through the sub-network responsible for outputting the connection position in the multi-task network, and output the connection travel through the sub-network responsible for outputting the connection travel mode. Way.

In this embodiment, in the process of processing the route calculation request, the number of connecting positions between the starting and ending points and the types of connecting travel modes are not limited. Generally speaking, when the distance between the starting and ending points is short, the number of confirmed connecting positions between the starting and ending points is less, and the types of connecting travel modes are less; when the distance between the starting and ending points is longer , there are more connection positions between the determined starting and ending points, and there are more types of connection travel modes.

As an example, for the route calculation request corresponding to the starting point A and the ending point B, it is determined that the connection positions between the starting and ending points include C and D. Between position points A and C, the travel mode is subway; between position points C and D, the travel mode is bus; between position points D and B, the travel mode is cycling.

Continuing to refer to FIG. 3 , FIG. 3 is a schematic diagram 300 of an application scenario of the method for determining travel information according to this embodiment. In the application scenario of FIG. 3 , the user 301 sends a route calculation request to the server 303 through the terminal device 302 . The route calculation request includes a start point 304 and an end point 305 . After determining the starting and ending points in the obtained road calculation request, the server first determines the road network time series features of the road network data 307, 308, 309 at multiple times within the time period to which the current moment belongs through the pre-trained feature extraction network 306 Then, the coding information corresponding to the starting and ending points is determined from the road network timing feature information; finally, based on the coding information, the pre-trained multi-task network 310 is used to determine the connection positions 311, 312 and the connection between the starting and ending points. Travel information for travel mode. Specifically, between location points 304 and 311 , the travel mode is subway; between 311 and 312 , the travel mode is bus; between 312 and 305 , the travel mode is cycling.

In this embodiment, a method for determining travel information is provided. For the road network time series feature information of the road network data at multiple times extracted by the feature extraction network in the travel information determination model, the time series corresponding to the starting point is determined. Accurate coding information, and then determining the connection position and connection travel mode between the starting and ending points based on the determined coding information through the multi-task network, which improves the accuracy of the obtained travel information and the flexibility of the travel mode.

In some optional implementations of this embodiment, the feature extraction network includes multiple graph convolutional networks and time series feature extraction networks. In this implementation manner, the above-mentioned execution subject may perform the above-mentioned step 202 in the following manner:

First, the road network feature information of road network data at multiple times in a time period is extracted through multiple graph convolutional networks.

As an example, there is a one-to-one correspondence between multiple graph convolutional networks and multiple times. For each graph convolutional network in the multiple graph convolutional networks, the road network data at the corresponding moment of the graph convolutional network is input into the graph convolutional network. network to obtain the road network characteristic information of the road network data at the corresponding time.

Second, the time sequence change information between the road network feature information corresponding to multiple moments in the time period is extracted through the time sequence feature extraction network, and the road network time sequence feature information is obtained.

In this implementation manner, after obtaining the road network feature information of the road network data at multiple times, the time sequence change information between the plurality of road network feature information is extracted through the time sequence feature extraction network to obtain the road network time sequence feature information.

The time series feature extraction network can be any network model with time series feature extraction function. As an example, the time series feature extraction network is an LSTM (long-short term memory, long short term memory) model with a GRU (Gated Recurrent Unit, Gated Recurrent Unit) structure.

In this implementation, firstly, the road network feature information of the road network data at each moment is extracted through multiple graph convolution networks, and then the road network time series feature information between the road network feature information corresponding to multiple moments is extracted through the time series feature extraction network. , the accuracy of the obtained road network time series feature information can be improved, and then the accuracy of the coding information of the starting and ending points can be improved, so as to improve the accuracy of the travel information.

In some optional implementations of this embodiment, the above-mentioned execution body may perform the above-mentioned step 204 in the following manner: based on the encoded information and the environmental information corresponding to the current moment, determine the connection position including the connection between the starting and ending points through the multi-task network and travel information for connecting modes of travel.

The environmental information may be any environmental information that can affect the travel mode, including but not limited to at least one of the following: weather, holidays, city size, preset activities, and seasons.

In this implementation manner, on the basis of considering the positions of the starting and ending points, the above-mentioned execution body also considers the environmental information at the current moment corresponding to the positions of the starting and ending points, so that under the influence of the environmental information, it generates an environment suitable for the current environment, including the starting and ending points. Travel information for the connecting location and connecting mode.

As an example, in a rainy day environment, the travel modes such as walking and cycling should be minimized; in a holiday environment, traffic congestion generally occurs, and the travel modes corresponding to the congested sections are reduced.

In this implementation, the travel information including the connection position between the start and end points and the connection travel mode is determined in combination with the coding information of the starting and ending points and the corresponding environmental information, so that the determined travel information is adapted to the environmental information, and further The accuracy of the determined travel information is improved.

Continuing to refer to FIG. 4 , a schematic structural diagram of the travel information determination model is shown. The travel information determination model 400 includes a feature extraction network 401 and a multi-task network 402 . The feature extraction network 401 includes a plurality of graph convolutional networks 4011 and a time series feature extraction network 4012 . The multiple graph convolutional networks 4011 respectively extract the road network feature information of the road network data at the corresponding time, and input the obtained road network feature information corresponding to the multiple times into the time series feature extraction network 4012 to obtain the road network time series feature information. The coding information of the starting and ending points in the road calculation request is determined from the road network time sequence feature information, and then the coding information and the environmental information are input into the multi-task network 402 to obtain the trip including the connecting position between the starting and ending points and the connecting travel mode. information.

In some optional implementation manners of this embodiment, the foregoing execution body may also perform the following operations:

First, the travel routes corresponding to the various connection travel modes are determined based on the starting and ending points and the connection locations through the graph search engine that has a one-to-one correspondence among the determined multiple connection travel modes.

In this implementation, the graph search engine is used to determine the travel route based on the road network data corresponding to the travel mode, including but not limited to the subway network data search engine, bus road network data search engine, driving road network data search engine, walking Road network data search engine, cycling road network data search engine.

After the connection location is determined, the endpoint information of multiple road sections between the starting and ending points in the route calculation request can be determined by combining the starting and ending points. The graph search engine determines the travel route for the road segment.

In some optional implementation manners, the above-mentioned executive body may determine the correspondence of various connection and travel modes based on the starting and ending points, connection positions and preset indicators through a graph search engine that has a one-to-one correspondence of the determined multiple connection and travel modes. travel route.

The preset indicators include, but are not limited to, indicators such as travel time, cost, and distance.

Under each index, for each road segment, the above-mentioned executive body may determine at least one travel route under the travel mode corresponding to the road segment. As an example, under the indicator of travel time, for the road segment between the starting point and the connecting position, the corresponding travel mode is public transportation, and the above-mentioned executive entity can determine the two bus travel routes that take the least time.

Second, a total travel route is obtained based on the determined travel routes corresponding to various connection travel modes.

As an example, the above-mentioned executive body may splicing travel routes corresponding to various connection travel modes according to the travel sequence between the starting and ending points to obtain the total travel route.

When the above-mentioned execution entity determines the travel routes corresponding to various connection travel modes based on the starting and ending points, connection positions and preset indicators, the various connection travel modes under the preset indicators can be spliced under each preset indicator. The corresponding travel route is obtained, and the total travel route corresponding to the preset index is obtained.

After obtaining the total travel route, the above executive body may display at least one obtained total travel route to the user for the user to select, and navigate to the user based on the selected total travel route.

In this implementation manner, after obtaining the connection location and the connection travel mode, the above-mentioned executive body determines the travel route of the corresponding road section through the graph search engine corresponding to the connection travel mode, so as to obtain the total travel route, which improves the accuracy of the travel route. Spend.

Continuing to refer to FIG. 5 , a schematic flow 500 of another embodiment of the method for determining travel information according to the present disclosure is shown, including the following steps:

Step 501: Determine the start and end points in the obtained route calculation request.

Step 502: Extract road network feature information of road network data at multiple times within a time period through multiple graph convolutional networks.

Step 503 , extracting time sequence variation information between road network feature information corresponding to multiple moments in a time period through a time sequence feature extraction network, to obtain road network time sequence feature information.

Step 504: Determine the coding information corresponding to the start and end points from the road network time sequence feature information.

Step 505 , based on the encoded information and the environmental information corresponding to the current moment, determine travel information including the connection location between the starting and ending points and the connection travel mode through the multi-task network.

Step 506: Determine travel routes corresponding to various connection travel modes based on the starting and ending points and the connection positions through a graph search engine that has a one-to-one correspondence of the determined multiple connection travel modes.

Step 507 , based on the determined travel routes corresponding to various connection travel modes, obtain a total travel route.

It can be seen from this embodiment that, compared with the embodiment corresponding to FIG. 2 , the flow 500 of the method for determining travel information in this embodiment specifically describes the process of determining the time series feature information of the road network, the process of determining travel information, And the process of determining the travel route improves the accuracy of the determined travel route.

Continuing to refer to FIG. 6 , a schematic flow 600 of an embodiment of a training method for a travel information determination model according to the present disclosure is shown, including the following steps:

Step 601, acquiring a training sample set.

In this embodiment, the execution body of the training method for the travel information determination model (for example, the terminal device or server in FIG. 1 ) may obtain a training sample set remotely or locally based on a wired network connection or a wireless network connection.

Among them, the training samples in the training sample set include trajectory data containing different travel modes, connection location labels, and connection travel mode labels.

As an example, the above-mentioned execution body may collect the user's trajectory data based on the user's mobile device, extract the trajectory data mixed with different travel modes, and determine the connection location and the connection travel mode from the trajectory data mixed with different travel modes, as a label.

Step 602: Determine the road network time sequence feature information of the road network data at multiple moments in the target time period corresponding to the input track data through the feature extraction network; determine the start and end points of the input track data from the road network time sequence feature information Corresponding coding information; input the coding information into the multi-task network, take the connection position label and connection travel mode label corresponding to the input trajectory data as the expected output of the multi-task network, and train to obtain the feature extraction network and multi-task network. Trip information determines the model.

In this embodiment, the above-mentioned execution body can determine, through the feature extraction network, the road network time series feature information of the road network data at multiple times within the target time period corresponding to the input trajectory data; The coding information corresponding to the starting and ending points of the trajectory data; input the coding information into the multi-task network, and use the connection position label and connection travel mode label corresponding to the input trajectory data as the expected output of the multi-task network, and the training results include feature extraction. Travel information determination model for network and multi-task network.

Road network data is road data under various travel modes, including but not limited to subway network data, bus road network data, driving road network data, walking road network data, and cycling road network data. Due to factors such as road construction, road control, and tidal lanes, road network data may change at different times. In order to determine the impact of changes in network data on travel information, the above-mentioned executive body can extract the feature information of road network data at multiple moments near the current moment through the feature extraction network, and then extract the time series changes of the feature information of road network data at each moment. information to obtain the road network timing characteristic information.

In this embodiment, the feature extraction network may be an encoding network with a feature extraction function, including but not limited to a convolutional neural network model, a residual network model, an AlexNet model, and a VGG (Visual Geometry Group Network) model. .

The target time period corresponding to the input trajectory data is the time period to which the generation time of the trajectory data belongs. The above-mentioned execution body may take days as a period, and divide a day into multiple times (for example, divide a day into 96 times on average). According to the time length corresponding to the preset time period as a time window, a plurality of times adjacent to the current time are determined. The length of the time period can be set flexibly. As an example, the duration of the time period is 120 minutes.

The road network time series feature information includes coded information of each position in the road network. In this embodiment, the above-mentioned execution body may determine the coding information of the starting point position and the coding information of the ending point position from the road network time series feature information based on the matching of the position information.

Input the coding information of the starting and ending points into the multi-task network, output the actual connection position through the sub-network responsible for outputting the connection position in the multi-task network, and output the actual connection and travel mode through the sub-network responsible for outputting the connection travel mode; and then determine the actual connection and travel mode. The first loss between the connection location and the connection location label, and the second loss between the actual connection travel mode and the connection travel mode label; The second loss determines the total loss, and the gradient is determined according to the total loss to update the parameters of the feature extraction network and the multi-task network.

By cyclically executing the above training operations until the preset end condition is reached, a trained travel information determination model is obtained. The preset end condition may be, for example, that the number of training times exceeds the preset number of times, the training time exceeds the preset time threshold, and the training loss tends to converge. The trained travel information determination model may be used to perform the method in the above-described embodiment 200 .

In this embodiment, a training method for a travel information determination model is provided. For the road network time series feature information of the road network data at multiple times extracted by the feature extraction network in the travel information determination model, it is determined from the The accurate encoding information of the starting point, and then the determined encoding information is used as the input of the multi-task network, and the connection position label and connection travel mode label corresponding to the input trajectory data are used as the expected output of the multi-task network to determine the start and end points. The connection location and connection travel mode between them improve the accuracy of the travel information determination model obtained by training.

In some optional implementations of this embodiment, the feature extraction network includes multiple graph convolutional networks and time-series feature extraction networks. In this implementation manner, the above-mentioned execution body may execute the following manner to determine, through the feature extraction network, the road network time series feature information of the road network data at multiple moments within the target time period corresponding to the input trajectory data:

First, the road network feature information of road network data at multiple times in a time period is extracted through multiple graph convolutional networks.

As an example, there is a one-to-one correspondence between multiple graph convolutional networks and multiple times. For each graph convolutional network in the multiple graph convolutional networks, the road network data at the corresponding moment of the graph convolutional network is input into the graph convolutional network. network to obtain the road network characteristic information of the road network data at the corresponding time.

Second, the time series change information between the road network characteristic information corresponding to multiple times is extracted through the time series feature extraction network, and the road network time series characteristic information is obtained.

In this implementation manner, after obtaining the road network feature information of the road network data at multiple times, the time sequence change information between the plurality of road network feature information is extracted through the time sequence feature extraction network to obtain the road network time sequence feature information.

The time series feature extraction network can be any network model with time series feature extraction function. As an example, the time series feature extraction network is an LSTM (long-short term memory, long short term memory) model with a GRU (Gated Recurrent Unit, Gated Recurrent Unit) structure.

In this implementation, firstly, the road network feature information of the road network data at each moment is extracted through multiple graph convolution networks, and then the road network time series feature information between the road network feature information corresponding to multiple moments is extracted through the time series feature extraction network. , the accuracy of the obtained road network time series feature information can be improved, and then the accuracy of the coding information of the starting and ending points can be improved, so as to improve the accuracy of the travel information.

In some optional implementation manners of this embodiment, the above-mentioned execution body may execute the following manner to input the encoded information into the multi-task network, and use the connection position label and connection travel mode label corresponding to the input trajectory data as multi-task Expected output of the network:

The encoded information and the environmental information corresponding to the generation time of the input trajectory data are input into the multi-task network, and the connection location label and the connection travel mode label corresponding to the input trajectory data are used as the expected output of the multi-task network.

The environmental information includes, but is not limited to, at least one of the following: weather, holidays, city size, preset activities, and seasons.

In this implementation manner, on the basis of considering the positions of the starting and ending points, the above-mentioned execution body also considers the environmental information at the current moment corresponding to the positions of the starting and ending points, so that under the influence of the environmental information, it generates an environment suitable for the current environment, including the starting and ending points. Travel information for the connecting location and connecting mode.

In this implementation, the travel information including the connection position between the start and end points and the connection travel mode is determined in combination with the coding information of the starting and ending points and the corresponding environmental information, so that the determined travel information is adapted to the environmental information, and further The accuracy of the determined travel information is improved.

Continuing to refer to FIG. 7 , as an implementation of the methods shown in the above figures, the present disclosure provides an embodiment of an apparatus for determining travel information. The apparatus embodiment corresponds to the method embodiment shown in FIG. 2 . The apparatus Specifically, it can be applied to various electronic devices.

As shown in FIG. 7 , the device for determining travel information includes: a first determining unit 701, configured to determine the starting and ending points in the acquired route calculation request; an extraction unit 702, configured to determine the current Road network timing feature information of the road network data at multiple moments in the time period to which the moment belongs; the coding unit 703 is configured to determine the coding information corresponding to the starting and ending points from the road network timing feature information; the second determining unit 704, is configured to determine travel information including connection locations between start and end points and connection travel modes through a pre-trained multi-task network based on the encoded information.

In some optional implementations of this embodiment, the feature extraction network includes multiple graph convolutional networks and time-series feature extraction networks, and the extraction unit 702 is further configured to: extract time periods respectively through multiple graph convolutional networks The road network feature information of the road network data at multiple times in the time period is extracted through the time series feature extraction network.

In some optional implementations of this embodiment, the second determining unit 704 is further configured to: determine the connection position including the connection position between the start and end points and the Travel information for connecting modes of travel.

In some optional implementation manners of this embodiment, the above-mentioned apparatus further includes: a third determination unit (not shown in the figure), configured to be a graph search engine that corresponds one-to-one through the determined multiple connection and travel modes , determine the travel routes corresponding to various connection travel modes based on the starting and ending points and the connection positions; the obtaining unit (not shown in the figure) is configured to obtain the total travel routes based on the determined travel routes corresponding to the various connection travel modes. travel route.

In this embodiment, a device for determining travel information is provided, which determines the time series feature information of the road network from the road network data at multiple times extracted by the feature extraction network in the travel information determination model, and determines the time series corresponding to the starting point. Accurate coding information, and then determining the connection position and connection travel mode between the starting and ending points based on the determined coding information through the multi-task network, which improves the accuracy of the obtained travel information and the flexibility of the travel mode.

Continuing to refer to FIG. 8 , as an implementation of the methods shown in the above figures, the present disclosure provides an embodiment of a training device for a travel information determination model, and the device embodiment corresponds to the method embodiment shown in FIG. 6 , Specifically, the device can be applied to various electronic devices.

As shown in FIG. 8 , the device for determining travel information includes: an obtaining unit 801 configured to obtain a training sample set, wherein the training samples in the training sample set include trajectory data containing different travel modes, connection location labels, connection travel The training unit 802 is configured to determine, through the feature extraction network, the road network time series feature information of the road network data at multiple moments in the target time period corresponding to the input trajectory data; The coding information corresponding to the starting and ending points of the input trajectory data; the coding information is input into the multi-task network, and the connection position label and connection travel mode label corresponding to the input trajectory data are used as the expected output of the multi-task network. Extraction network, travel information determination model of multi-task network.

In some optional implementations of this embodiment, the feature extraction network includes multiple graph convolutional networks and time-series feature extraction networks; and the training unit 802 is further configured to: extract time periods respectively through multiple graph convolutional networks The road network feature information of the road network data at multiple times in the system is extracted; the time series change information between the road network feature information corresponding to the multiple times is extracted through the time series feature extraction network, and the road network time series feature information is obtained.

In some optional implementations of this embodiment, the training unit 802 is further configured to: input the encoded information and the environmental information corresponding to the generation time of the input trajectory data into the multi-task network, and input the input trajectory data into the multi-task network. The corresponding connection location label and connection travel mode label are used as the expected output of the multi-task network.

In this embodiment, a training device for a travel information determination model is provided. For the road network time series feature information of the road network data at multiple times extracted by the feature extraction network in the travel information determination model The accurate coding information of the starting point, and then the determined coding information is the input of the multi-task network, and the connection position label and the connection travel mode label corresponding to the input trajectory data are the expected output of the multi-task network, and the starting and ending points are determined. The connection location and the connection travel mode between the trains improve the accuracy of the travel information determination model obtained by training.

According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, the electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores data executable by the at least one processor. The instruction is executed by at least one processor, so that when executed by the at least one processor, the method for determining travel information and the method for training a model for determining travel information described in any of the foregoing embodiments can be implemented.

According to an embodiment of the present disclosure, the present disclosure also provides a readable storage medium, where the readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to implement the travel information described in any of the above embodiments when executed. The determination method and the travel information determine the training method of the model.

Embodiments of the present disclosure provide a computer program product, which, when executed by a processor, can implement the method for determining travel information and the method for training a model for determining travel information described in any of the foregoing embodiments.

FIG. 9 shows a schematic block diagram of an example electronic device 900 that may be used to implement embodiments of the present disclosure. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic devices may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are by way of example only, and are not intended to limit implementations of the disclosure described and/or claimed herein.

As shown in FIG. 9 , the device 900 includes a computing unit 901 that can be executed according to a computer program stored in a read only memory (ROM) 902 or a computer program loaded from a storage unit 908 into a random access memory (RAM) 903 Various appropriate actions and handling. In the RAM 903, various programs and data necessary for the operation of the device 900 can also be stored. The computing unit 901 , the ROM 902 , and the RAM 903 are connected to each other through a bus 904 . An input/output (I/O) interface 905 is also connected to bus 904 .

Various components in the device 900 are connected to the I/O interface 905, including: an input unit 906, such as a keyboard, mouse, etc.; an output unit 907, such as various types of displays, speakers, etc.; a storage unit 908, such as a magnetic disk, an optical disk, etc. ; and a communication unit 909, such as a network card, a modem, a wireless communication transceiver, and the like. The communication unit 909 allows the device 900 to exchange information/data with other devices through a computer network such as the Internet and/or various telecommunication networks.

Computing unit 901 may be various general-purpose and/or special-purpose processing components with processing and computing capabilities. Some examples of computing units 901 include, but are not limited to, central processing units (CPUs), graphics processing units (GPUs), various specialized artificial intelligence (AI) computing chips, various computing units that run machine learning model algorithms, digital signal processing processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 901 executes the various methods and processes described above, such as the determination method of travel information. For example, in some embodiments, the method of determining travel information may be implemented as a computer software program tangibly embodied on a machine-readable medium, such as storage unit 908 . In some embodiments, part or all of the computer program may be loaded and/or installed on device 900 via ROM 902 and/or communication unit 909 . When the computer program is loaded into RAM 903 and executed by computing unit 901, one or more steps of the above-described determination method of travel information may be performed. Alternatively, in other embodiments, the computing unit 901 may be configured to perform the determination method of travel information by any other suitable means (eg, by means of firmware).

Various implementations of the systems and techniques described herein above may be implemented in digital electronic circuitry, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips system (SOC), load programmable logic device (CPLD), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include being implemented in one or more computer programs executable and/or interpretable on a programmable system including at least one programmable processor that The processor, which may be a special purpose or general-purpose programmable processor, may receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device an output device.

Program code for implementing the methods of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general purpose computer, special purpose computer or other programmable data processing apparatus, such that the program code, when executed by the processor or controller, performs the functions/functions specified in the flowcharts and/or block diagrams. Action is implemented. The program code may execute entirely on the machine, partly on the machine, partly on the machine and partly on a remote machine as a stand-alone software package or entirely on the remote machine or server.

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

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

The systems and techniques described herein may be implemented on a computing system that includes back-end components (eg, as a data server), or a computing system that includes middleware components (eg, an application server), or a computing system that includes front-end components (eg, a user's computer having a graphical user interface or web browser through which a user may interact with implementations of the systems and techniques described herein), or including such backend components, middleware components, Or any combination of front-end components in a computing system. The components of the system may be interconnected by any form or medium of digital data communication (eg, a communication network). Examples of communication networks include: Local Area Networks (LANs), Wide Area Networks (WANs), and the Internet.

A computer system can include clients and servers. Clients and servers are generally remote from each other and usually interact through a communication network. The relationship of client and server arises by computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also known as a cloud computing server or a cloud host. It is a host product in the cloud computing service system to solve the management difficulties in traditional physical host and virtual private server (VPS, Virtual Private Server) services. The defect is large and weak in business scalability; it can also be a server of a distributed system, or a server combined with blockchain.

According to the technical solutions of the embodiments of the present disclosure, a method for determining travel information is provided, which is determined by using the road network time series feature information of road network data at multiple times extracted by a feature extraction network in a travel information determination model, and corresponding to Accurate coding information of the starting point, and then determining the connecting position and connecting travel mode between the starting and ending points based on the determined coding information through the multi-task network, which improves the accuracy of the obtained travel information and the flexibility of the travel mode.

It should be understood that steps may be reordered, added or deleted using the various forms of flow shown above. For example, the steps described in the present disclosure can be executed in parallel, sequentially, or in different orders. As long as the desired results of the technical solutions provided in the present disclosure can be achieved, no limitation is imposed herein.

The above-mentioned specific embodiments do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may occur depending on design requirements and other factors. Any modifications, equivalent replacements, and improvements made within the spirit and principles of the present disclosure should be included within the protection scope of the present disclosure.

Claims (17)

1. A travel information determination method comprises the following steps:

determining a starting point and a stopping point in the acquired route calculation request;

determining road network time sequence characteristic information of road network data at multiple moments in a time period to which the current moment belongs through a pre-trained characteristic extraction network;

determining coding information corresponding to the starting point and the ending point from the road network timing characteristic information;

and determining trip information comprising a connection position between the starting point and the stopping point and a connection trip mode through a pre-trained multi-task network based on the coding information.

2. The method of claim 1, wherein the feature extraction network comprises a plurality of graph convolution networks and a time-series feature extraction network, an

The method for determining road network time sequence feature information of road network data at multiple moments in a time period to which the current moment belongs through the pre-trained feature extraction network comprises the following steps:

respectively extracting road network characteristic information of road network data at a plurality of moments in the time period through the plurality of graph convolution networks;

and extracting time sequence change information among the road network characteristic information corresponding to a plurality of moments in the time period through the time sequence characteristic extraction network to obtain the road network time sequence characteristic information.

3. The method of claim 1, wherein said determining, based on said encoded information, trip information comprising a docked position and a docked trip mode between said start and stop points over a pre-trained multi-tasking network comprises:

and determining trip information including a connection position between the starting point and the stopping point and a connection trip mode through the multi-task network based on the coded information and the environment information corresponding to the current moment.

4. The method according to any one of claims 1-3, further comprising:

determining travel routes corresponding to various connection travel modes on the basis of the starting and stopping points and the connection positions through the determined graph search engines corresponding to the various connection travel modes one by one;

and obtaining a total travel route based on the determined travel routes corresponding to the various connection travel modes.

5. A training method of a travel information determination model comprises the following steps:

acquiring a training sample set, wherein training samples in the training sample set comprise track data containing different trip modes, connection position labels and connection trip mode labels;

determining road network time sequence characteristic information of road network data at a plurality of moments in a target time period corresponding to the input track data through a characteristic extraction network; determining coding information corresponding to the start point and the end point of the input track data from the road network time sequence characteristic information; inputting the coding information into a multi-task network, taking a connection position label and a connection travel mode label corresponding to the input track data as expected output of the multi-task network, and training to obtain a travel information determination model comprising the feature extraction network and the multi-task network.

6. The method of claim 5, wherein the feature extraction network comprises a plurality of graph convolution networks and a time-series feature extraction network; and

the method for determining road network time series characteristic information of road network data at a plurality of moments in a target time period corresponding to input trajectory data through a characteristic extraction network comprises the following steps:

respectively extracting road network characteristic information of road network data at a plurality of moments in the time period through the plurality of graph volume networks;

and extracting time sequence change information among the road network characteristic information corresponding to a plurality of moments through the time sequence characteristic extraction network to obtain the road network time sequence characteristic information.

7. The method of claim 5, wherein the inputting the encoded information into a multitasking network, and the outputting the docking position tag and the docking travel mode tag corresponding to the input trajectory data as the expected output of the multitasking network comprises:

and inputting the coding information and the environment information corresponding to the generation moment of the input track data into a multi-task network, and taking a connection position label and a connection travel mode label corresponding to the input track data as expected output of the multi-task network.

8. A travel information determination apparatus comprising:

a first determination unit configured to determine a start point and a stop point in the acquired route calculation request;

the extraction unit is configured to determine road network time sequence characteristic information of road network data at a plurality of moments in a time period to which the current moment belongs through a pre-trained characteristic extraction network;

the encoding unit is configured to determine encoding information corresponding to the start point and the stop point from the road network timing characteristic information;

a second determining unit configured to determine, based on the encoded information, travel information including a connection position and a connection travel pattern between the start and stop points through a pre-trained multitask network.

9. The apparatus of claim 8, wherein the feature extraction network comprises a plurality of graph convolution networks and a time-series feature extraction network, an

The extraction unit, further configured to:

respectively extracting road network characteristic information of road network data at a plurality of moments in the time period through the plurality of graph volume networks; and extracting time sequence change information among the road network characteristic information corresponding to a plurality of moments in the time period through the time sequence characteristic extraction network to obtain the road network time sequence characteristic information.

10. The apparatus of claim 8, wherein the second determining unit is further configured to:

and determining trip information including a connection position between the starting point and the stopping point and a connection trip mode through the multi-task network based on the coded information and the environment information corresponding to the current moment.

11. The apparatus of any of claims 8-10, further comprising:

a third determining unit configured to determine, by using the graph search engine in which the determined multiple connection travel modes correspond to one another, travel routes corresponding to the various connection travel modes based on the start and stop points and the connection positions;

and the obtaining unit is configured to obtain a total travel route based on the travel routes corresponding to the determined various connection travel modes.

12. A training apparatus for a travel information determination model, comprising:

the system comprises an acquisition unit, a processing unit and a processing unit, wherein the acquisition unit is configured to acquire a training sample set, wherein training samples in the training sample set comprise track data containing different travel modes, connection position labels and connection travel mode labels;

a training unit configured to determine, by a feature extraction network, road network timing feature information of road network data at a plurality of times within a target time period corresponding to the input trajectory data; determining coding information corresponding to the start point and the end point of the input track data from the road network time sequence characteristic information; inputting the coding information into a multi-task network, taking a connection position label and a connection travel mode label corresponding to the input track data as expected output of the multi-task network, and training to obtain a travel information determination model comprising the feature extraction network and the multi-task network.

13. The apparatus of claim 12, wherein the feature extraction network comprises a plurality of graph convolution networks and a time-series feature extraction network; and

the training unit, further configured to:

respectively extracting road network characteristic information of road network data at a plurality of moments in the time period through the plurality of graph convolution networks; and extracting time sequence change information among the road network characteristic information corresponding to a plurality of moments through the time sequence characteristic extraction network to obtain the road network time sequence characteristic information.

14. The apparatus of claim 12, wherein the training unit is further configured to:

and inputting the coding information and the environment information corresponding to the generation moment of the input track data into a multi-task network, and taking a connection position label and a connection travel mode label corresponding to the input track data as expected output of the multi-task network.

15. An electronic device, comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-7.

16. A non-transitory computer readable storage medium having stored thereon computer instructions for causing the computer to perform the method of any one of claims 1-7.

17. A computer program product, comprising: computer program which, when being executed by a processor, carries out the method according to any one of claims 1-7.

Images