Navigation route recommendation, familiar road prediction model training method, device and medium

Abstract

The present disclosure provides a method and device for training a familiar route prediction model, and a medium for recommending a navigation route, wherein the method comprises: obtaining historical travel data of a candidate navigation route including a road segment; constructing prediction data of the road segment according to the historical travel data of the road segment and route planning data of the candidate navigation route, the prediction data including planning feature data of the road segment; and determining a recommended prediction value of the candidate navigation route based on the prediction data of the candidate navigation route including the road segment. The technical solution of the present disclosure can timely capture the travel preference changes of a navigation object, more accurately recommend a navigation route meeting the preference of the navigation object, and better achieve personalized route recommendation.

Description

Technical Field

[0001] This disclosure relates to the field of navigation technology, and in particular to a method, apparatus, and medium for recommending navigation routes and training a familiar route prediction model. Background Technology

[0002] Existing applications (software) and their service systems that support map navigation can calculate and recommend navigation routes for users based on their input of starting point, destination, and other information.

[0003] To ensure that recommended navigation routes meet users' travel preferences (such as a preference for familiar routes or routes with fewer traffic lights), existing technologies typically rely on data representing these preferences to determine the recommended routes. However, the inventors have discovered that users' travel preferences change over time and with varying environments. When the data representing these preferences fails to reflect these changes, the recommended routes may not meet user expectations, resulting in a poor user experience. Summary of the Invention

[0004] This disclosure provides a method, apparatus, and medium for recommending navigation routes and training a familiar route prediction model.

[0005] In a first aspect, embodiments of this disclosure provide a method for recommending navigation routes, including:

[0006] Obtain historical travel data for candidate navigation routes, including road segments;

[0007] Based on historical travel data of the road segment and route planning data of candidate navigation routes, predictive data for the road segment is constructed, including planning feature data of the road segment.

[0008] Based on the predicted data of the candidate navigation routes including road segments, the recommended predicted values ​​of the candidate navigation routes are determined.

[0009] Secondly, embodiments of this disclosure provide a method for training a familiar route prediction model, comprising:

[0010] Obtain sample navigation routes, including historical travel data for sample road segments;

[0011] Based on the sample historical travel data of the sample road segments and the sample route planning data of the sample navigation routes, training samples of the sample road segments are constructed. The training samples include the sample planning feature data of the sample road segments.

[0012] The familiar route prediction model is trained using training samples from sample road segments. The familiar route prediction model is used to determine the recommended prediction value of the candidate navigation route based on the prediction data of the candidate navigation route, including road segments.

[0013] Thirdly, embodiments of this disclosure provide a navigation route recommendation device, comprising:

[0014] The first acquisition module is used to acquire historical travel data of candidate navigation routes, including road segments.

[0015] The first construction module is used to construct the prediction data of the road segment based on the historical travel data of the road segment and the route planning data of the candidate navigation routes. The prediction data includes the planning feature data of the road segment.

[0016] The determination module is used to determine the recommended predicted value of the candidate navigation route based on the predicted data of the road segments included in the candidate navigation route.

[0017] Fourthly, embodiments of this disclosure provide a training apparatus for a familiar route prediction model, comprising:

[0018] The second acquisition module is used to acquire sample navigation routes, including sample road segments, and sample historical travel data.

[0019] The second construction module is used to construct training samples for the sample road segments based on the sample historical travel data of the sample road segments and the sample route planning data of the sample navigation routes. The training samples include the sample planning feature data of the sample road segments.

[0020] The training module is used to train the familiar route prediction model using training samples from sample road segments. The familiar route prediction model is used to determine the recommended prediction value of the candidate navigation route based on the prediction data of the candidate navigation route, including road segments.

[0021] Fifthly, embodiments of this disclosure provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method provided in any embodiment of this disclosure.

[0022] The technical solution disclosed herein involves acquiring historical travel data for candidate navigation routes, including road segments; constructing predicted data for road segments based on the historical travel data and route planning data of the candidate navigation routes, the predicted data including planning feature data of the road segments; and determining recommended predicted values ​​for candidate navigation routes based on the predicted data for road segments. Since the historical travel data in this disclosure is generated from the actual travel behavior of the navigated object, it characterizes the navigated object's actual travel preferences for the corresponding road segments. When the navigated object's travel preferences change with time or environmental factors, the historical travel data can easily capture these changes in preference. Furthermore, since the predicted data of the road segments constructed by this disclosure based on historical travel data and route planning data of candidate navigation routes includes the planning feature data of the road segments, that is, the predicted data of the road segments constructed by this disclosure associates the features of the road segments included in the candidate navigation routes with the features recorded in the historical travel data of the road segments. This makes the changes in the travel preferences of the navigated object reflected in the predicted data of the candidate navigation routes, thereby improving the accuracy of the recommended predicted values ​​of the candidate routes. This makes the navigation routes recommended to the navigated object more in line with the current travel preferences of the navigated object, and the navigated object is more likely to choose the recommended navigation routes.

[0023] The above overview is for illustrative purposes only and is not intended to be limiting in any way. Further aspects, embodiments, and features of this disclosure will become readily apparent from the accompanying drawings and the following detailed description, in addition to the illustrative aspects, embodiments, and features described above. Attached Figure Description

[0024] In the accompanying drawings, unless otherwise specified, the same reference numerals throughout the various drawings denote the same or similar parts or elements. These drawings are not necessarily drawn to scale. It should be understood that these drawings depict only some embodiments disclosed in this disclosure and should not be construed as limiting the scope of this disclosure.

[0025] Figure 1 This is a flowchart illustrating a navigation route recommendation method in one embodiment of the present disclosure;

[0026] Figure 2 This is a flowchart illustrating the training method of a familiar route prediction model in one embodiment of the present disclosure;

[0027] Figure 3 This is a structural block diagram of a navigation route recommendation device according to an embodiment of the present disclosure;

[0028] Figure 4 This is a structural block diagram of a training device for a familiar route prediction model in one embodiment of the present disclosure;

[0029] Figure 5 This is a block diagram of an electronic device used to implement embodiments of the present disclosure. Detailed Implementation

[0030] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of this disclosure. Therefore, the drawings and description are to be considered exemplary in nature and not restrictive.

[0031] To facilitate understanding of the technical solutions of the embodiments of this disclosure, the related technologies of the embodiments of this disclosure are described below. The following related technologies are optional solutions and can be combined with the technical solutions of the embodiments of this disclosure in any way, and they all fall within the protection scope of the embodiments of this disclosure.

[0032] The user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this disclosure are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data shall comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points shall be provided for users to choose to authorize or refuse.

[0033] When using a map navigation application, the user (also known as the navigated object) inputs the starting point and destination. The application and its related service systems then plan the optimal navigation route and alternative routes under different strategies. Furthermore, to ensure that the recommended navigation route meets the user's expectations, existing technologies also incorporate the user's travel preferences to recommend routes that align with their expectations; for example, recommending routes consisting of familiar road segments.

[0034] Some of the concepts involved in this disclosure include road segments and personal road networks.

[0035] Link: The smallest unit of a road. It can be understood that a road in the real world is divided into multiple links in an electronic map. Usually, a section of road between two intersections is divided into a link. Because roads are divided into links in electronic maps, navigation routes planned based on electronic maps usually include more than one link.

[0036] Personal road network: It is obtained by mining the road information that the user (also known as the navigation object) has actually walked (driving or walking) within a preset historical time period (e.g., within six months). The personal road network is recorded according to road segments.

[0037] In related technologies, a personal road network represents the routes a user prefers or is familiar with when traveling. Existing technologies typically determine personal road networks using predefined rules, such as classifying routes a user has traveled twice or more as preferred / familiar routes. However, personal road networks mined using predefined rules have some problems. For example, defining a route as familiar only as one traveled twice or more can lead to changes in the environment over time; these routes may no longer be frequently used. In other words, over time, these routes may no longer be familiar or preferred by the user, resulting in navigation routes that do not necessarily reflect the user's desired route. Especially in the familiar route mode, the navigation route that the user expects may not be the same as the route that the personal road network is usually developed over a long period of time. When the user changes their work location, moves, or has some special needs, and the route they used to frequently take is no longer the route they currently travel on, the personal road network will still mark this route as a familiar route because it is not updated in time. This will cause the predicted value of the navigation route that includes this route to usually exceed that of other navigation routes, and ultimately the navigation route recommended to the user will not actually meet the user's current expectations.

[0038] Therefore, there is a need for new navigation route recommendation technologies that can capture changes in users' travel preferences in a timely manner and recommend navigation routes that meet users' expectations.

[0039] Figure 1 This is a flowchart illustrating a navigation route recommendation method according to an embodiment of this disclosure. Figure 1 As shown, the navigation route recommendation method 100 provided in this embodiment includes steps S101 to S103.

[0040] In step S101, historical travel data of the candidate navigation routes, including road segments, are obtained.

[0041] The number of candidate navigation routes can be one or more, and the candidate navigation routes can include several topologically connected road segments. The historical travel data of the road segments can include historical feature data characterizing the historical travel preferences of the navigated object. The historical travel data obtained in step S101 can be the historical travel data of the road segments within a preset historical time range, for example, within six months or three months from the current time. The preset historical time range can be set as needed, and this disclosure does not impose any restrictions.

[0042] Historical travel data corresponds to the object being navigated. For example, historical travel data for the corresponding road segment can be obtained based on the identifier of the object being navigated. This historical travel data records historical characteristic data generated when the object being navigated actually traversed the corresponding road segment before the route planning time of the candidate navigation route. Because the object being navigated had actually traversed this road segment before, and because the historical travel data is relatively recent, the historical characteristic data used in this disclosure can characterize recent changes in the object being navigated's historical travel preferences for this road segment.

[0043] The road segments in the candidate navigation routes may or may not have historical travel data. For example, if the navigated object passed through a road segment within a preset historical time range, then that road segment has corresponding historical travel data; if the navigated object did not pass through that road segment within the preset historical time range, then that road segment has no historical travel data. This disclosure pertains to road segments in the candidate navigation routes for which historical travel data is available. For road segments for which historical travel data is not available, the calculation of their weights can employ relevant existing technologies, and this disclosure does not impose any restrictions.

[0044] In step S102, prediction data for the road segment is constructed based on historical travel data and route planning data of candidate navigation routes. The prediction data includes planning feature data of the road segment.

[0045] In step S102, the route planning data can be understood as the data required by applications and service systems with map navigation functions when planning candidate navigation routes, including but not limited to the starting point, destination, and road conditions. Based on historical travel data of the road segments and the route planning data of the candidate navigation routes, predictive data for the road segments is constructed. This predictive data includes planning feature data for the road segments. This planning feature data correlates the features of the road segments included in the candidate navigation routes during planning with the features recorded in the historical travel data of those road segments, thus reflecting changes in the travel preferences of the navigated individual in the predictive data of the candidate navigation routes.

[0046] In step S103, based on the predicted data of the road segments included in the candidate navigation route, the recommended predicted value of the candidate navigation route is determined.

[0047] The recommended prediction value refers to the score given to the navigator by the candidate navigation route. The higher the recommended prediction value, the higher the probability that the candidate navigation route will be recommended to the navigator. When the recommendation scenario is "familiar route mode," the recommended prediction value can be understood as the navigator's familiarity with the candidate navigation route. The higher the recommended prediction value, the higher the navigator's familiarity with the candidate navigation route. For the navigator, especially those driving, driving on familiar routes provides a clearer understanding of road conditions and greater peace of mind. Therefore, recommending familiar navigation routes reduces the navigator's sense of unfamiliarity with the route, better meets their expectations, and allows them to choose familiar routes, thus improving driving safety.

[0048] The above describes the method provided by this disclosure. This method determines the recommended predicted value of each candidate navigation route. After determining the recommended predicted value of each candidate navigation route, this method compares the recommended predicted value of the candidate navigation route with a preset threshold. From the candidate navigation routes whose recommended predicted value is greater than the preset threshold, the method selects the candidate navigation routes recommended to the navigated object. For example, when there are at least two candidate navigation routes whose recommended predicted value is greater than or equal to the preset threshold, these candidate navigation routes can be sorted according to the size of their recommended predicted values. Based on the sorting, one or more candidate navigation routes can be recommended as target navigation routes to the navigated object. The number of candidate navigation routes recommended to the navigated object can be set as needed, and this disclosure does not impose any limitations.

[0049] In related technologies, personal road networks are typically determined using predefined rules. For example, routes traveled twice or more are considered preferred or familiar to the user. However, personal road networks derived from predefined rules have several issues. While routes traveled twice or more may be considered familiar, they can change over time or due to environmental changes. This means that routes containing such routes may not reflect the user's desired route. Especially in the "familiar route" mode, due to the long development cycle of personal road networks, outdated networks may still mark routes as familiar. This results in routes containing such routes often having higher predicted values ​​than other routes, failing to accurately reflect the user's travel preferences. Ultimately, the recommended routes may not match the user's current expectations.

[0050] The technical solution disclosed herein involves acquiring historical travel data for candidate navigation routes, including road segments; constructing predicted data for road segments based on the historical travel data and route planning data of the candidate navigation routes, the predicted data including planning feature data of the road segments; and determining recommended predicted values ​​for candidate navigation routes based on the predicted data for road segments. Since the historical travel data in this disclosure is generated from the actual travel behavior of the navigated object, it characterizes the navigated object's actual travel preferences for the corresponding road segments. When the navigated object's travel preferences change with time or environmental factors, the historical travel data can easily capture these changes in preference. Furthermore, since the predicted data for road segments constructed in this disclosure based on historical travel data and route planning data of candidate navigation routes includes the planning feature data of the road segments, the predicted data for road segments constructed in this disclosure correlates the features of the road segments included in the candidate navigation routes during planning with the features recorded in the historical travel data. This ensures that changes in the travel preferences of the navigated user are reflected in the predicted data of the candidate navigation routes, thereby improving the accuracy of the recommended predicted values ​​of the candidate routes. This makes the navigation routes recommended to the navigated user more consistent with their current travel preferences, and the navigated user is more likely to choose the recommended navigation routes. Therefore, by adopting the technical solution of this disclosure, changes in the travel preferences of the navigated user can be captured in a timely manner, and navigation routes that match their preferences can be recommended more accurately, thus better achieving personalized route recommendations.

[0051] In one embodiment, historical travel data may include historical travel time, which can be understood as the historical time when the navigated object traversed the road segment. Route planning data may include route planning time. Route planning time should be understood as the time taken by a navigation application and its service system to plan a navigation route based on the starting point and destination input by the navigated object. The planning characteristic data of the road segment includes time characteristics and distance characteristics.

[0052] Based on historical travel data and route planning data for candidate navigation routes, predictive data for the road segment can be constructed, which may include:

[0053] Based on the historical travel time included in the historical travel data of the road segment and the route planning time included in the route planning data of the candidate navigation routes, determine the time characteristics corresponding to the historical travel time;

[0054] Based on the historical travel data of the road segment, the historical origin and destination corresponding to the historical travel time, and the planned origin and destination included in the route planning data of the candidate navigation route, determine the distance characteristics corresponding to the historical travel time; at least based on the time characteristics and distance characteristics, construct the prediction data of the road segment.

[0055] Understandably, the greater the difference between the historical travel time and the route planning time for a road segment, the more time has passed since the navigated user previously traveled that segment and the planned candidate navigation route. Furthermore, the navigated user's familiarity with a road segment decreases over time; that is, the greater the difference between the historical travel time and the route planning time, the less familiar the navigated user is with that road segment. Conversely, if the historical travel time and the route planning time are very close, the navigated user is more familiar with that road segment. Therefore, this disclosure, by determining the time characteristics corresponding to the historical travel time of a road segment based on the historical travel time and the route planning time of the candidate navigation route, considers the issue of familiarity decaying over time, making the constructed prediction data more reflective of the navigated user's preference for or familiarity with the corresponding road segment.

[0056] The historical start point and historical end point can be the start and end points of the historical travel route corresponding to the historical travel time. The historical travel route is the navigation route corresponding to when the navigated object previously passed through this road segment. For example, if the navigated object planned a navigation route from start point E to end point F 5 days before the route planning time, passing through road segment 1, and the navigated object actually traveled through road segment 1, then the historical start point and historical end point corresponding to the historical travel time are respectively the start and end points of the navigation route from start point E to end point F. The historical travel time can be the specific date of passing through the historical travel route, such as November 20, 2022, or it can be the number of days between the time of passing through the historical travel route and the route planning time, such as 5 days ago. The specific format of the historical travel time can be set as needed. Continuing the previous example, if the navigated object passed through road segment 1 5 days before the candidate navigation route planning time, and the historical travel route through road segment 1 is from start point E to end point F, then for road segment 1, the historical feature data can include 5 days ago, start point E, and end point F, and the historical travel data can be represented as {5 days ago, start point E, end point F}. Each time the navigated object passes through road segment 1, a historical feature data point corresponding to road segment 1 is generated. The more times the object passes through road segment 1, the more historical feature data is generated for road segment 1. The planning start point and planning end point are the start and end points of the planned candidate navigation routes.

[0057] In related technologies, for familiar road segments of the navigated object, the starting and ending points of the navigation route are not considered. As long as the road segment exists in the personal road network, the probability of recommending a navigation route containing that road segment to the navigated object increases. In this embodiment, historical travel data includes the historical starting and ending points corresponding to historical travel times, and route planning data includes the planned starting and ending points. It can be understood that if the distance between the historical starting point and the planned starting point of the navigation route when the navigated object has previously traveled a certain road segment differs greatly, and the distance between the historical ending point and the planned ending point also differs greatly, it indicates that even if the navigated object has traveled that road segment before, it is not familiar with that road segment when traveling from the planned starting point to the planned ending point, or does not prefer that road segment. Conversely, if the distance between the historical starting point and the planned starting point of the navigation route when the navigated object has previously traveled a certain road segment is not significantly different, and the distance between the historical ending point and the planned ending point also differs little, it indicates that the navigated object prefers to travel that road segment when traveling from the planned starting point to the planned ending point.

[0058] In the embodiments described above, the time characteristics corresponding to historical travel times are determined based on historical travel times and route planning times; the distance characteristics corresponding to historical travel times are determined based on historical starting points, historical ending points, planned starting points, and planned ending points; and prediction data for road segments is constructed based at least on these time and distance characteristics. It is evident that the prediction data for road segments constructed in this disclosure not only considers the impact of time on travel preferences but also the impact of starting points and ending points on travel preferences, which can further improve the accuracy of the recommended prediction values ​​for candidate navigation routes, enabling the recommended prediction values ​​to more accurately reflect the current preferences of the navigated object for the candidate navigation routes.

[0059] It should be noted that the starting point in this disclosure can be the coordinates of the starting point, and the ending point can be the coordinates of the ending point. When the distance between two starting points is less than or equal to a preset distance, the two starting points can be considered the same; when the distance between two ending points is less than or equal to a preset distance, the two ending points can be considered the same. The preset distance can be, for example, 100m.

[0060] In one embodiment, the time feature may include the time difference between historical travel time and route planning time.

[0061] In one embodiment, the distance feature may include a first distance feature and a second distance feature. Based on the historical origin and destination corresponding to historical travel times in the historical travel data of the road segment, and the planned origin and destination included in the route planning data of the candidate navigation route, the distance feature corresponding to the historical travel time is determined, including: determining the distance between the historical origin and the planned origin as the first distance feature; and determining the distance between the historical destination and the planned destination as the second distance feature.

[0062] In the predicted data for road segments, the time difference between historical travel time and route planning time can be represented as the difference in the number of days between the historical travel time and the route planning time. This time difference characterizes the difference between the historical travel time of the navigated user passing through that road segment and the route planning time. The smaller the time difference, the closer the navigated user's time passing through that road segment is to the route planning time, indicating that the navigated user is more familiar with that road segment, or that the navigated user prefers that road segment. Therefore, the predicted data for road segments that includes the time difference between historical travel time and route planning time takes into account the impact of the time difference on travel preferences, making the recommended predicted values ​​of the determined candidate navigation routes more accurately reflect the navigated user's current preference for the candidate navigation routes.

[0063] The recommended predicted value of a candidate navigation route is related to the first distance feature and the second distance feature. Generally, the smaller the values ​​corresponding to the first and second distance features, the greater the likelihood that the navigated object will choose that route segment. Therefore, the predicted data for a route segment includes both the first and second distance features. The recommended predicted value of the candidate navigation route determined based on the predicted data of the route segment can better reflect the navigated object's familiarity with the route segment, better reflect the navigated object's current preference for the candidate navigation route, and better meet the navigated object's expectations.

[0064] Since road conditions determine the travel time cost of a navigation route, the target user typically does not actively choose frequently congested road segments. To further improve the accuracy of the predicted values ​​for candidate navigation routes and to reflect the target user's preference for road segments due to road conditions during the recommendation prediction process, in one embodiment of this disclosure, the road segment planning feature data may further include road condition features. The method in this embodiment may also include:

[0065] Based on the historical travel data of the road segment, the historical travel time corresponding to the historical traffic information, and the planned traffic information of the road segment included in the route planning data, the traffic characteristics corresponding to the historical travel time are determined.

[0066] For example, traffic information may include smooth traffic, slow traffic, congested traffic, extremely congested traffic, no traffic conditions, or traffic under construction. For instance, if historical traffic information indicates smooth traffic but planned traffic information indicates congested traffic, then the target user's current preference for the candidate navigation route is low, and the likelihood of the target user selecting that candidate navigation route is relatively small. It should be noted that "no traffic conditions" means that the traffic conditions for the road segment are unavailable.

[0067] In one embodiment, traffic condition characteristics may include the difference between historical traffic condition information corresponding to historical travel times and planned traffic condition information for the road segment. For example, each type of traffic condition information can be assigned a value: smooth traffic is 1, slow traffic is 0.8, congested traffic is 0.6, extremely congested traffic is 0.4, no traffic is 0.2, and under construction is 0. The difference between historical and planned traffic condition information can be determined as the traffic condition characteristic.

[0068] In one embodiment, traffic characteristics may include historical traffic information corresponding to historical travel times and planned traffic information for road segments included in the route planning data.

[0069] It should be noted that historical traffic information and planned traffic information are corresponding. When the historical traffic information corresponding to the historical travel time is the historical traffic information of a road segment, the planned traffic information is the traffic information of that road segment in the candidate navigation route at the planned time. When the historical traffic information corresponding to the historical travel time is the overall historical traffic information of the historical travel route, the planned traffic information is the overall traffic information of the candidate navigation route at the planned time.

[0070] Furthermore, since the greater the proportion of a familiar road segment in the candidate navigation route, the more the navigable object prefers that route, in one embodiment of this disclosure, the road segment planning feature data also includes a road segment proportion feature. The method in this embodiment may further include: determining the road segment proportion feature based on the proportion of the road segment's length in the candidate navigation route.

[0071] The segment proportion feature can reflect the navigation object's preference for the navigation route. When the segment prediction data includes the segment proportion feature, the recommended prediction value of the candidate navigation route determined based on the segment prediction data can further improve the accuracy of the recommended prediction value and better reflect the current preference of the navigation object for the candidate navigation route.

[0072] In this embodiment of the disclosure, prediction data for road segments is constructed based at least on time and distance features, including:

[0073] The time characteristics, distance characteristics, road condition characteristics, and road segment proportion characteristics corresponding to the same historical travel time are used to construct a prediction data for a road segment.

[0074] Understandably, each time the navigated object passes through a road segment, a corresponding historical travel data point is generated. Each historical travel data point corresponds to a historical travel time, and each historical travel time corresponds to a prediction data point. A prediction data point for a road segment can be represented as {time feature, distance feature, traffic condition feature, road segment percentage feature}. Alternatively, a prediction data point can be represented as {time feature, first distance feature, second distance feature, historical traffic condition information, planned traffic condition information, road segment percentage feature}.

[0075] In one embodiment, determining a recommended predicted value for a candidate navigation route based on predicted data of road segments included in the candidate navigation route includes: inputting the predicted data of road segments included in the candidate navigation route into a trained familiar route prediction model to obtain the model prediction value corresponding to the predicted data of the road segments; determining the road segment prediction value based on the model prediction value corresponding to the predicted data of the road segments; and determining the recommended predicted value for the candidate navigation route based on the road segment prediction value included in the candidate navigation route.

[0076] Each historical travel time for a road segment corresponds to a predicted data point for that segment. By inputting the predicted data of the road segment into a trained familiar road prediction model, the model prediction value corresponding to the predicted data of the road segment can be obtained.

[0077] The familiar route prediction model can be a neural network model, either a single-layer or multi-layer neural network. Training the familiar route prediction model allows it to learn the travel preferences of the navigated object at different times and with different origins and destinations. Since the predicted data for road segments is constructed based on historical travel data and route planning data of candidate navigation routes, the predicted data for road segments includes the planning feature data of the road segments. Historical travel data does not require special mining; the travel data of the navigated object itself is generated. Therefore, after the model is trained, inputting the predicted data of road segments into the trained familiar route prediction model allows the model's output prediction value to better reflect the current preference of the navigated object for that road segment. This avoids the problem of inaccurate recommended prediction values ​​caused by using personal road networks mined by predefined rules that cannot capture changes in the travel preferences of the navigated object in a timely manner.

[0078] In one embodiment, a candidate navigation route may include multiple road segments, each of which may correspond to at least one historical travel data point, and thus each road segment may correspond to at least one predicted data point. Therefore, the recommended predicted value of the candidate navigation route can be determined based on the predicted data corresponding to each road segment in the candidate navigation route.

[0079] Based on the model prediction values ​​corresponding to the prediction data of the road segment, determine the road segment prediction value, including: weighting the model prediction values ​​corresponding to the prediction data of the road segment to obtain the road segment prediction value.

[0080] Each historical travel time for a road segment corresponds to one predicted data point for that segment. When the navigated object passes through the road segment multiple times, multiple historical travel times exist for that segment, allowing the construction of multiple predicted data points. Correspondingly, the road segment corresponds to multiple model prediction values. The model prediction values ​​corresponding to the road segment can be weighted and calculated to obtain the road segment's predicted value.

[0081] For example, at least based on the time characteristics in the prediction data, determine the weight values ​​of the model predictions corresponding to the prediction data, and calculate the product of the model predictions and the weight values. Add the products of the model predictions to obtain the road segment prediction value.

[0082] For example, the model prediction value of a road segment is y. Assume candidate navigation route A includes road segments 1, 2, and 3. The model prediction values ​​for the three road segments corresponding to road segment 1 are y1, y2, and y3, respectively. Assume that the weight values ​​for y1, y2, and y3 are determined as α1, α2, and α3 based on the time characteristics in the prediction data. Then, the predicted value of road segment 1 is M1 = y1*α1 + y2*α2 + y3*α3. It should be noted that determining the weight values ​​for the model prediction values ​​corresponding to the prediction data is not limited to the time characteristics in the prediction data; it can also be based on distance characteristics, etc.

[0083] In one embodiment, the weight values ​​of the model prediction values ​​can all be set to 1, and the predicted value of road segment 1 is M1 = y1 + y2 + y3.

[0084] The road segment prediction value determined in this embodiment of the disclosure integrates the historical travel data of the navigated object each time it passes through the road segment, which can better reflect the navigated object's preference for that road segment. It is understood that the more times the navigated object passes through the road segment, the stronger its preference for that road segment. Therefore, by weighting the prediction values ​​of each model, the larger the calculated value, the larger the road segment prediction value, the stronger the navigated object's preference for that road segment, and the more familiar the navigated object is with that road segment.

[0085] In one embodiment, determining the recommended prediction value of a candidate navigation route based on the predicted values ​​of road segments included in the candidate navigation route includes: weighting the predicted values ​​of road segments included in the candidate navigation route to obtain the recommended prediction value of the candidate navigation route.

[0086] For example, the weight of a road segment's predicted value can be determined at least based on its proportion of length in the candidate navigation route. The product of the predicted road segment value and its weight value is then calculated. The products of the predicted road segments are summed to obtain the recommended predicted value for the candidate navigation route.

[0087] For example, the predicted value for road segment 2 is calculated to be M2. Road segment 3 has no historical travel data, therefore, road segment 3 has no predicted value or its predicted value is 0. The weights of M1 and M2 are determined to be β1 and β2, respectively. Then, the recommended predicted value for candidate navigation route 1 is Q1 = M1*β1 + M2*β2. It should be noted that determining the weights of road segment predicted values ​​is not limited to the length proportion of the road segment in the candidate navigation route; other relevant features can also be used to determine the weights of the road segment predicted values.

[0088] In one embodiment, the weight values ​​of the predicted road segment values ​​can all be set to 1, and the recommended predicted value of candidate navigation route A is Q1 = M1 + M2.

[0089] This embodiment of the disclosure determines the recommended prediction value of a candidate navigation route by considering all road segments historically traversed by the navigated object within the candidate navigation route. The more road segments historically traversed by the navigated object within the candidate navigation route, the more familiar the navigated object is with that candidate navigation route. Therefore, the higher the recommended prediction value of the obtained candidate navigation route, the more familiar the navigated object is with that candidate navigation route, and the more likely the navigated object is to choose that candidate navigation route.

[0090] In this embodiment, the predicted data for a road segment includes the planned feature data of the road segment, which includes time features, a first distance feature, a second distance feature, historical traffic information, planned traffic information, and a road segment percentage feature. Therefore, the predicted data for a road segment can be represented as {time features, first distance features, second distance features, historical traffic information, planned traffic information, and road segment percentage feature}. The predicted data for a road segment can have 6 dimensions. The predicted data for a road segment in this embodiment has strong scalability. If a new road segment feature is added, it is only necessary to expand the dimensions of the predicted data for the road segment from 6 dimensions to 7 dimensions, thus enhancing the scalability of the technical solution disclosed in this invention.

[0091] In one embodiment, the navigation route recommendation method can be applied to the server side. Before obtaining historical travel data of road segments for candidate navigation routes, the navigation route recommendation method may further include: obtaining a planned starting point and a planned ending point, and planning at least one candidate navigation route based on the planned starting point and the planned ending point. For example, after the navigable object enters the planned starting point and planned ending point in an application with map navigation functionality, the client can send the planned starting point, planned ending point, and client identifier to the server, where the client identifier corresponds to the navigable object. The server plans at least one candidate navigation route based on the obtained planned starting point and planned ending point, whereby the candidate navigation route is a navigation route from the planned starting point to the planned ending point. The server obtains historical travel data of road segments for the candidate navigation route based on the client identifier.

[0092] The server determines the recommended navigation route to the navigated object based on the predicted values ​​of candidate navigation routes. If the predicted values ​​of all candidate navigation routes are less than a preset threshold, the server may choose not to recommend any of the candidate routes to the navigated object. The server may also use other methods to determine and recommend the navigation route to the navigated object.

[0093] In one embodiment, the navigation route recommendation method can be applied to the server side. Before obtaining historical travel data of the road segments in the candidate navigation routes, the navigation route recommendation method may further include: obtaining candidate navigation routes. For example, after the navigable object on the client side inputs the planned starting point and planned ending point in an application that supports map navigation, the client application plans candidate navigation routes based on the planned starting point and planned ending point, and sends the candidate navigation routes and client identifiers to the server side. The server side obtains the historical travel data of the road segments in the candidate navigation routes based on the obtained candidate navigation routes and client identifiers. The historical travel data corresponds to the navigable object, and the navigable object corresponds to the client identifier.

[0094] It should be noted that when the navigating object uses an application that supports map navigation, the system can record the various road segments traversed by the navigating object, the time taken, and the corresponding start and end points of its historical travel routes. This data is stored and can be considered the historical travel data of the navigating object. The historical travel data corresponds to the navigating object, and different navigating objects may have different historical travel data.

[0095] In other embodiments, when the starting point and ending point of other navigated objects are the same as the planned starting point and planned ending point, in order to improve the efficiency of navigation route recommendation, the navigation route recommended to the navigated object can be directly recommended to the other navigated object, or the route with the highest recommended prediction value among the navigation routes recommended to the navigated object can be recommended to the other navigated object.

[0096] In one embodiment, at least one candidate navigation route can be recommended based on the recommended predicted value of the candidate navigation route and the attribute information of the candidate navigation route. The attribute information of the candidate navigation route may include the total length of the candidate navigation route, the number of traffic lights on the candidate navigation route, the current time of passing through the candidate navigation route, etc.

[0097] It's important to note that the factors influencing the selection of candidate navigation routes by the navigator include not only the navigator's familiarity with the route but also the attribute information of the candidate navigation routes. Understandably, if a candidate navigation route has a high predicted recommendation value, but the actual travel time along that route far exceeds the navigator's expectations, the navigator may not choose that route. Therefore, recommending at least one candidate navigation route based on its predicted recommendation value and attribute information, which is not only a route the navigator is familiar with but also considers the actual conditions at the time of route planning, can further improve the navigator's satisfaction.

[0098] The following specific embodiments illustrate the application process of the navigation route recommendation method disclosed herein.

[0099] For example, candidate navigation routes include route A, route B, and route C. The segment sequence of each route is extracted to obtain the segment sequences of the three routes: route A {segment 1, segment 2, segment 3}; route B {segment 4, segment 5, segment 6, segment 7}; route C {segment 1, segment 2, segment 8, segment 7}.

[0100] For route A, iterate through each segment of route A. If a segment lacks historical travel data, it is ignored. For example, segment 1 lacks historical travel data, while segments 2 and 3 do. Therefore, segment 1 is ignored, and the historical travel data for segments 2 and 3 is obtained. Based on the historical travel data for segment 2 and the route planning data for route A, predictive data for segment 2 in route A is constructed. The predicted data for segment 2 is input into a trained familiar route prediction model to obtain the model prediction value corresponding to the predicted data for segment 2. The model prediction values ​​for segment 2 are weighted to obtain the predicted value for segment 2. Similarly, the predicted value for segment 3 is obtained. The predicted values ​​for segments 2 and 3 are weighted to obtain the recommended predicted value for route A.

[0101] Using the same method, we obtained the recommended predicted values ​​for routes B and C.

[0102] Suppose that the recommended prediction value for route A is 45, for route B it is 50, and for route C it is 65, with a preset threshold of 50. The recommended prediction values ​​for routes A, B, and C are compared with the preset threshold, and routes with a recommended prediction value greater than or equal to the threshold are identified as the target navigation routes. The resulting target navigation routes include routes B and C. Routes B and C are then sorted according to their recommended prediction values. The route with the highest recommended prediction value among the target navigation routes can be recommended to the navigable object; for example, route C can be recommended to the navigable object.

[0103] Figure 2 This is a flowchart illustrating a training method for a familiar route prediction model in one embodiment of this disclosure. This disclosure also provides a training method 200 for a familiar route prediction model, such as... Figure 2 As shown, the training method 200 includes steps S201 to S203.

[0104] In step S201, sample navigation routes, including sample road segments, are obtained from historical travel data.

[0105] The number of sample navigation routes can be one or more. A sample navigation route may include several topologically connected sample road segments. The sample historical travel data for each sample road segment may include sample historical feature data characterizing the historical travel preferences of the navigated object. The sample historical travel data can be sample historical travel data for sample road segments within a preset historical time range, for example, within six months or three months of the sample planning time. The preset historical time range can be set as needed.

[0106] Historical travel data corresponds to the object being navigated. For example, based on the identifier of the object being navigated, historical travel data for the corresponding sample road segment can be obtained. This historical travel data records the historical characteristics of the sample road segment actually traversed before the planned route time of the sample navigation route. Because the object being navigated had previously traversed this sample road segment, the historical characteristics of the sample road segment can characterize recent changes in the object's historical travel preferences for that segment.

[0107] The sample road segments in the sample navigation route may or may not have historical travel data. For example, if the navigated object passes through the sample road segment within a preset historical time range, then that sample road segment has corresponding historical travel data; if the navigated object does not pass through the sample road segment within the preset historical time range, then that sample road segment has no historical travel data. This disclosure pertains to sample road segments in the sample navigation route for which historical travel data is available. For sample road segments for which historical travel data is not available, the calculation of their weights can employ relevant existing technologies, and this disclosure does not impose any restrictions.

[0108] In step S202, training samples for the sample road segments are constructed based on the sample historical travel data of the sample road segments and the sample route planning data of the sample navigation routes. The training samples include the sample planning feature data of the sample road segments.

[0109] In step S202, the sample route planning data can be understood as the data required by applications and service systems with map navigation functions when planning sample navigation routes, including but not limited to sample start point, sample end point, and sample road conditions. Training samples for sample road segments are constructed based on the sample historical travel data of the sample road segments and the sample route planning data of the sample navigation routes. The training samples include sample planning feature data of the sample road segments. This data correlates the features of the sample road segments included in the sample navigation routes during planning with the features recorded in the sample historical travel data, thus reflecting changes in the travel preferences of the navigated user in the predicted data of the sample navigation routes.

[0110] In step 203, the familiar route prediction model is trained using training samples of the sample road segments. The familiar route prediction model is used to determine the recommended prediction value of the candidate navigation route based on the prediction data of the road segments included in the candidate navigation route.

[0111] The training method provided in this disclosure involves acquiring historical travel data of sample navigation routes, including sample road segments; constructing training samples of sample road segments based on the historical travel data of the sample road segments and the sample route planning data of the sample navigation routes, wherein the training samples include sample planning feature data of the sample road segments; and training the familiar route prediction model using the training samples of the sample road segments. Since the historical travel data is generated from the actual travel behavior of the navigated object, it characterizes the navigated object's actual travel preferences for the corresponding sample road segments. When the navigated object's travel preferences change with time or environment, the historical travel data can easily capture these changes in preference. Furthermore, since this disclosure constructs training samples of sample road segments based on historical travel data and sample route planning data of sample navigation routes, meaning that the training samples of sample road segments constructed by this disclosure associate the features of the sample road segments included in the sample navigation route during planning with the features recorded in the historical travel data of the sample road segments, the changes in the travel preferences of the navigated object are reflected in the training samples of the sample road segments. Therefore, after training the familiar route prediction model using the training samples determined by the embodiments of this disclosure, the model prediction values ​​output by the familiar route prediction model can more accurately reflect the preferences of the navigated object for the sample road segments at the time of sample navigation route planning. The recommended prediction values ​​of candidate navigation routes determined by the model prediction values ​​output by the familiar route prediction model can more accurately reflect the current travel preferences of the navigated object for the candidate navigation routes. When a navigation route is recommended to the navigated object based on the recommended prediction values, the recommended navigation route is more in line with the current travel preferences of the navigated object, and the navigated object is more likely to choose the recommended navigation route. Therefore, the familiar route prediction model trained using the technical solution disclosed herein can capture changes in the travel preferences of the navigated object in a timely manner, and can more accurately recommend navigation routes that match the navigated object's preferences, thus better realizing personalized route recommendations.

[0112] It should be noted that when the familiar route prediction model is used in the navigation route recommendation method of this embodiment, the input of the familiar route prediction model is the prediction data of road segments. Therefore, the training samples of the sample road segments can be understood according to the relevant content of the prediction data of the road segments.

[0113] In one embodiment, sample historical travel data may include sample historical travel time, which can be understood as the historical time when the sample being navigated traversed the sample road segment. Sample route planning data may include sample route planning time. Sample route planning time should be understood as the time it takes for a navigation application and its service system to plan a sample navigation route based on the sample starting point and sample ending point input by the sample being navigated.

[0114] The sample planning feature data of the sample road segment includes: sample time features and sample distance features. Based on the sample historical travel data of the sample road segment and the sample route planning data of the sample navigation route, training samples of the sample road segment are constructed. This may include: determining the sample time features corresponding to the sample historical travel time based on the sample historical travel time included in the sample historical travel data of the sample road segment and the sample route planning time included in the sample route planning data of the sample navigation route; determining the sample distance features corresponding to the sample historical travel time based on the sample historical start and end points corresponding to the sample historical travel time in the sample historical travel data of the sample road segment, and the sample planning start and end points included in the sample route planning data of the sample navigation route; training samples of the sample road segment are constructed based on at least the sample time features and sample distance features.

[0115] Understandably, the greater the difference between the historical travel time and the route planning time for a sample road segment, the more distant the time since the route planning time has been. Since the familiarity of the navigated user with the sample road segment decreases over time, the greater the difference between the historical travel time and the route planning time, the less familiar the navigated user is with the sample road segment. Conversely, a close difference between the historical travel time and the route planning time indicates greater familiarity with the sample road segment. Therefore, the sample time feature determined based on the historical travel time of the sample road segment and the route planning time of the sample navigation route, taking into account the issue of familiarity decaying over time, better reflects the navigated user's preference for or familiarity with the corresponding sample road segment.

[0116] The sample historical start point and sample historical end point can be the start and end points of the sample historical travel routes corresponding to the sample historical travel times. The sample historical travel routes are the navigation routes corresponding to when the sample was navigated through the sample road segment in the past. The sample planning start point and sample planning end point are the start and end points of the sample navigation routes.

[0117] In this embodiment, the sample historical travel data includes the sample historical starting point and sample historical ending point corresponding to the sample historical travel time, and the sample route planning data includes the sample planning starting point and sample planning ending point. It is understood that if the distance between the sample historical starting point and the sample planning starting point, and the distance between the sample historical ending point and the sample planning ending point, of the navigation route when the navigated object has historically traversed a certain sample road segment, differs significantly, it indicates that even if the navigated object has previously traversed that sample road segment, it is not familiar with that sample road segment or does not prefer it when traveling from the sample planning starting point to the sample planning ending point. Conversely, if the distance between the sample historical starting point and the sample planning starting point, and the distance between the sample historical ending point and the sample planning ending point, of the navigation route when the navigated object has historically traversed a certain sample road segment, are not significantly different, it indicates that the navigated object prefers that sample road segment when traveling from the sample planning starting point to the sample planning ending point.

[0118] In the embodiments of this disclosure, sample time features corresponding to sample historical travel times are determined based on sample historical travel times and sample route planning times; sample distance features corresponding to sample historical travel times are determined based on sample historical starting points, sample historical ending points, sample planning starting points, and sample planning ending points; and training samples of sample road segments are constructed based on at least these sample time features and sample distance features. It is evident that the training samples of sample road segments constructed in this disclosure consider both the impact of time on the travel preferences of the navigated object and the impact of the sample planning starting point and sample planning ending point on travel preferences. Therefore, the familiar route prediction model trained using these training samples considers the impact of time on travel preferences, as well as the impact of the route starting point and ending point on travel preferences. Consequently, the recommended prediction values ​​of candidate navigation routes determined using the familiar route prediction model of this disclosure can more accurately reflect the current preferences of the navigated object for candidate navigation routes, better reflect the familiarity of the navigated object with navigation routes, and more accurately recommend familiar routes to the navigated object, thus better achieving personalized route recommendations.

[0119] In one embodiment, the sample time feature may include the time difference between the sample's historical travel time and the sample's route planning time.

[0120] In one embodiment, the sample distance features include a first sample distance feature and a second sample distance feature. Based on the sample historical travel data of the sample road segment, which corresponds to the sample historical travel time, and the sample planning start point and sample planning end point included in the sample route planning data of the sample navigation route, the sample distance features corresponding to the sample historical travel time are determined, including: determining the distance between the sample historical start point and the sample planning start point as the first sample distance feature; and determining the distance between the sample historical end point and the sample planning end point as the second sample distance feature.

[0121] It should be noted that the greater the difference between the historical travel data of the sample road segment and the route planning data of the sample navigation route, the greater the difference between the sample navigation route and the historical travel route. Correspondingly, the lower the familiarity of the navigated object with the sample navigation route; conversely, the higher the familiarity of the navigated object with the sample navigation route. Therefore, the distance between the historical starting point and the planned starting point of the sample is defined as the first distance feature, and the distance between the historical ending point and the planned ending point of the sample is defined as the second distance feature. This allows the training samples to better reflect the difference between the historical travel data of the sample road segment and the route planning data of the sample navigation route. After training the familiar route prediction model with these training samples, the accuracy of the familiar route prediction model's output can be improved, and the obtained recommended prediction values ​​can more accurately reflect the current preference of the navigated object for the planned navigation route.

[0122] In one embodiment, the sample planning feature data of the sample road segment further includes sample traffic condition features, and the method further includes: determining the sample traffic condition features corresponding to the sample historical travel time based on the sample historical travel data of the sample road segment and the sample planning traffic condition information of the sample road segment included in the sample route planning data.

[0123] Traffic information for a route or road segment can reflect, and even determine, whether the navigated object chooses that route or road segment. Therefore, traffic information can reflect the navigated object's preference for that route or road segment. Thus, by determining the sample traffic characteristics corresponding to the sample's historical travel time based on sample historical traffic information and sample planned traffic information, and by including these sample traffic characteristics in the training samples, the accuracy of the familiar route prediction model trained using these training samples can be further improved.

[0124] For example, traffic information may include smooth traffic, slow traffic, congested traffic, extremely congested traffic, no traffic conditions, or traffic under construction. For instance, if the historical traffic information for a sample is smooth traffic, but the planned traffic information for the sample is congested, then the target user has a low preference for the sample navigation route, and the likelihood of the target user choosing that route is relatively small. It should be noted that "no traffic conditions" means that the traffic conditions for the road segment are unknown.

[0125] In one embodiment, the sample traffic condition feature may include the difference between the sample historical traffic condition information corresponding to the sample historical travel time and the sample planned traffic condition information for the sample road segment. For example, each type of traffic condition information is assigned a value: smooth traffic is 1, slow traffic is 0.8, congested traffic is 0.6, extremely congested traffic is 0.4, no traffic is 0.2, and under construction is 0. The difference between the sample historical traffic condition information and the sample planned traffic condition information can be determined as the sample traffic condition feature.

[0126] In one embodiment, traffic characteristics may include historical traffic information corresponding to historical travel times of the samples and planned traffic information of the sample road segments included in the sample route planning data.

[0127] It should be noted that the sample historical traffic information and the sample planned traffic information are corresponding. When the sample historical travel time corresponds to the sample historical traffic information of the sample road segment, the sample planned traffic information is the traffic information of that sample road segment in the sample navigation route at the sample planning time. When the sample historical travel time corresponds to the sample historical traffic information of the overall historical traffic information of the sample historical travel route, the sample planned traffic information is the overall traffic information of the sample navigation route at the planning time.

[0128] In one embodiment, the sample planning feature data of the sample road segment also includes the sample road segment proportion feature, and the method further includes: determining the sample road segment proportion feature based on the length proportion of the sample road segment in the sample navigation route.

[0129] Understandably, the greater the proportion of sample road segments in the sample navigation route, the more the navigated object prefers that sample navigation route. Therefore, when the training samples include the proportion of sample road segments, using training samples that include this characteristic to train the familiar route prediction model can further improve the accuracy of the model's output, further improve the accuracy of the recommended prediction values ​​of the candidate navigation routes determined by the familiar route prediction model, and better reflect the current preference of the navigated object for the candidate navigation routes.

[0130] In one embodiment, training samples for sample road segments are constructed based at least on sample time features and sample distance features, including: constructing a training sample for a sample road segment by taking the sample time features, sample distance features, sample traffic condition features, and sample road segment proportion features corresponding to the historical travel time of the same sample.

[0131] It is understandable that each time the navigable object passes through a sample road segment, a corresponding sample historical travel data is generated. Each sample historical travel data corresponds to a sample historical travel time, and each sample historical travel time corresponds to a training sample. A training sample of a sample road segment can be represented as {sample time feature, sample distance feature, sample traffic condition feature, sample road segment percentage feature}. A training sample can also be represented as {sample time feature, sample first distance feature, sample second distance feature, sample historical traffic condition information, sample planned traffic condition information, sample road segment percentage feature}.

[0132] In one embodiment, training the familiar route prediction model using training samples of sample road segments includes: inputting the training samples of sample road segments into the familiar route prediction model to obtain the model prediction values ​​corresponding to the training samples of sample road segments; and adjusting the parameters of the familiar route prediction model based on the model prediction values ​​corresponding to the training samples of sample road segments, sample navigation routes, and sample travel routes.

[0133] In one embodiment, the parameters of the familiar route prediction model are adjusted based on the model prediction values ​​corresponding to the training samples of the sample road segments, the sample navigation routes, and the sample travel routes. This includes: determining the travel coverage rate of the sample navigation routes, where the travel coverage rate is the proportion of the length of overlapping road segments in the sample navigation routes; determining the sample recommended prediction values ​​of the sample navigation routes based on the model prediction values ​​corresponding to the training samples of the sample road segments; determining the loss value of the familiar route prediction model based on the sample recommended prediction values ​​and the travel coverage rate of the sample navigation routes; and adjusting the parameters of the familiar route prediction model based on the loss value.

[0134] In one embodiment, determining the sample recommendation prediction value of the sample navigation route based on the model prediction value corresponding to the training samples of the sample road segment includes: inputting the training samples of the sample navigation route, including the sample road segment, into a familiar route prediction model to obtain the model prediction value corresponding to the training samples of the sample road segment; determining the sample road segment prediction value of the sample road segment based on the model prediction value corresponding to the training samples of the sample road segment; and determining the sample recommendation prediction value of the sample navigation route based on the sample road segment prediction value of the sample navigation route, including the sample road segment.

[0135] In one embodiment, determining the sample road segment prediction value of a sample road segment based on the model prediction values ​​corresponding to the training samples of the sample road segment includes: weighting the model prediction values ​​corresponding to the training samples of the sample road segment to obtain the sample road segment prediction value of the sample road segment.

[0136] For example, based at least on the temporal features in the training samples, determine the weight values ​​of the model predictions corresponding to the training samples, and calculate the product of the model predictions and the weight values. Add the products of the model predictions to obtain the sample road segment predictions.

[0137] In one embodiment, the sum of the model predictions corresponding to the training samples of the sample road segment can be used as the sample road segment prediction value.

[0138] In one embodiment, determining the sample recommendation prediction value of a sample navigation route based on the sample segment prediction values ​​of the sample navigation route including sample road segments includes: weighting the sample segment prediction values ​​of the sample navigation route including sample road segments to obtain the sample recommendation prediction value of the sample navigation route.

[0139] For example, the weight value of the predicted value of the sample road segment can be determined at least based on the proportion of the sample road segment's length in the sample navigation route. The product of the predicted value of the sample road segment and the weight value is calculated. The products corresponding to the predicted values ​​of the sample road segments are added together to obtain the sample recommended predicted value of the sample navigation route.

[0140] In one embodiment, the sum of the predicted values ​​of sample road segments that include sample road segments in the sample navigation route can be used as the sample recommendation predicted value of the sample navigation route.

[0141] The sample recommendation prediction value of the sample navigation route is determined, and the travel coverage rate of the sample navigation route is used as a label to adjust the parameters of the familiar route prediction model.

[0142] For example, the training samples include sample time features, sample first distance features, sample second distance features, sample historical traffic information, sample planned traffic information, and sample road segment proportion features. The parameters of the familiar road prediction model include the weight values ​​of sample time features, sample first distance features, sample second distance features, sample historical traffic information, sample planned traffic information, and sample road segment proportion features. Different training samples are used to train the familiar road prediction model to obtain the weight values ​​of sample time features, sample first distance features, sample second distance features, sample historical traffic information, sample planned traffic information, and sample road segment proportion features that make the loss value converge.

[0143] In this embodiment, the training samples include sample time features, sample first distance features, sample second distance features, sample historical traffic information, sample planned traffic information, and sample road segment percentage features. The training samples can be represented as {sample time features, sample first distance features, sample second distance features, sample historical traffic information, sample planned traffic information, and sample road segment percentage features}. The training samples have a 6-dimensional dimension. The training samples in this embodiment have strong scalability; if a new sample planning feature is added, the dimensionality of the training samples only needs to be expanded from 6 dimensions to 7 dimensions, thus enhancing the scalability of the technical solution disclosed in this invention.

[0144] Figure 3 This is a structural block diagram of a navigation route recommendation device according to an embodiment of this disclosure. Figure 3 As shown, the navigation route recommendation device 300 includes a first acquisition module 301, a first construction module 302, and a determination module 303.

[0145] The first acquisition module 301 is used to acquire historical travel data of candidate navigation routes, including road segments.

[0146] The first construction module 302 is used to construct the prediction data of the road segment based on the historical travel data of the road segment and the route planning data of the candidate navigation routes. The prediction data includes the planning feature data of the road segment.

[0147] The determination module 303 is used to determine the recommended predicted value of the candidate navigation route based on the predicted data of the road segments included in the candidate navigation route.

[0148] In one embodiment, the planning feature data of the road segment includes: time features and distance features. The first construction module 302 includes: a first determining submodule, used to determine the time features corresponding to the historical travel time based on the historical travel time included in the historical travel data of the road segment and the route planning time included in the route planning data of the candidate navigation route; a second determining submodule, used to determine the distance features corresponding to the historical travel time based on the historical starting point and historical ending point corresponding to the historical travel time in the historical travel data of the road segment, and the planning starting point and planning ending point included in the route planning data of the candidate navigation route; and a construction submodule, used to construct the prediction data of the road segment based at least on the time features and distance features.

[0149] In one embodiment, the distance features include a first distance feature and a second distance feature, and the second determining submodule is used to: determine the distance between the historical starting point and the planned starting point as the first distance feature; and determine the distance between the historical ending point and the planned ending point as the second distance feature.

[0150] In one embodiment, the planning feature data of the road segment also includes road condition features, and the first construction module 302 further includes: a third determining submodule, used to determine the road condition features corresponding to the historical travel time based on the historical road condition information corresponding to the historical travel time in the historical travel data of the road segment, and the planned road condition information of the road segment included in the route planning data.

[0151] In one embodiment, the planning feature data of the road segment also includes the road segment proportion feature, and the first construction module 302 further includes: a fourth determination submodule, used to determine the road segment proportion feature based on the length proportion of the road segment in the candidate navigation route.

[0152] In one embodiment, the construction submodule is used to construct a prediction data for a road segment by combining the time characteristics, distance characteristics, road condition characteristics, and road segment proportion characteristics corresponding to the same historical travel time.

[0153] In one embodiment, the determining module is further configured to: input the predicted data of the candidate navigation route including road segments into a trained familiar route prediction model to obtain the model prediction value corresponding to the predicted data of the road segments; determine the road segment prediction value of the road segments based on the model prediction value corresponding to the predicted data of the road segments; and determine the recommended prediction value of the candidate navigation route based on the road segment prediction value of the candidate navigation route including road segments.

[0154] In one embodiment, the determining module is further configured to: perform weighted calculation on the model prediction values ​​corresponding to the prediction data of the road segment to obtain the road segment prediction value; and perform weighted calculation on the road segment prediction values ​​of the candidate navigation route to obtain the recommended prediction value of the candidate navigation route.

[0155] Figure 4 This is a structural block diagram of a training device for a familiar route prediction model in one embodiment of this disclosure. Figure 4 As shown, the training device 400 for the familiar route prediction model includes a second acquisition module 401, a second construction module 402, and a training module 403.

[0156] The second acquisition module 401 is used to acquire sample navigation routes, including sample road segments, of sample historical travel data.

[0157] The second construction module 402 is used to construct training samples for the sample road segments based on the sample historical travel data of the sample road segments and the sample route planning data of the sample navigation routes. The training samples include the sample planning feature data of the sample road segments.

[0158] The training module 403 is used to train the familiar route prediction model using training samples of sample road segments. The familiar route prediction model is used to determine the recommended prediction value of the candidate navigation route based on the prediction data of the candidate navigation route including road segments.

[0159] In one embodiment, the sample planning feature data of the sample road segment includes: sample time features and sample distance features. The second construction module 402 is further configured to: determine the sample time features corresponding to the sample historical travel time based on the sample historical travel time included in the sample historical travel data of the sample road segment and the sample route planning time included in the sample route planning data of the sample navigation route; determine the sample distance features corresponding to the sample historical travel time based on the sample historical start point and sample historical end point corresponding to the sample historical travel time in the sample historical travel data of the sample road segment, and the sample planning start point and sample planning end point included in the sample route planning data of the sample navigation route; and construct training samples of the sample road segment based at least on the sample time features and sample distance features.

[0160] In one embodiment, the sample distance features include a first sample distance feature and a second sample distance feature. The second construction module 402 is further configured to: determine the distance between the sample historical starting point and the sample planning starting point as the first sample distance feature; and determine the distance between the sample historical ending point and the sample planning ending point as the second sample distance feature.

[0161] In one embodiment, the sample planning feature data of the sample road segment also includes sample road condition features, and the second construction module 402 is further configured to: determine the sample road condition features corresponding to the sample historical travel time based on the sample historical travel data of the sample road segment and the sample planning road condition information of the sample road segment included in the sample route planning data.

[0162] In one embodiment, the sample planning feature data of the sample road segment also includes the sample road segment proportion feature, and the second construction module 402 is further used to: determine the sample road segment proportion feature based on the length proportion of the sample road segment in the sample navigation route.

[0163] In one embodiment, the second construction module 402 is further configured to: construct a training sample of a sample road segment from the sample time features, sample distance features, sample road condition features and sample road segment proportion features corresponding to the historical travel time of the same sample.

[0164] In one embodiment, the training module includes: an acquisition submodule, used to input training samples of sample road segments into the familiar route prediction model to obtain model prediction values ​​corresponding to the training samples of sample road segments; and an adjustment submodule, used to adjust the parameters of the familiar route prediction model based on the model prediction values ​​corresponding to the training samples of sample road segments, sample navigation routes, and sample travel routes.

[0165] In one embodiment, the adjustment submodule is used to: determine the travel coverage rate of the sample navigation route, where the travel coverage rate is the proportion of the length of overlapping segments of the sample navigation route and the sample travel route in the sample navigation route; determine the sample recommendation prediction value of the sample navigation route based on the model prediction value corresponding to the training samples of the sample segments; determine the loss value of the familiar route prediction model based on the sample recommendation prediction value and the travel coverage rate of the sample navigation route; and adjust the parameters of the familiar route prediction model based on the loss value.

[0166] Figure 5 This is a block diagram of an electronic device used to implement embodiments of the present disclosure. For example... Figure 5 As shown, the electronic device includes a memory 510 and a processor 520. The memory 510 stores a computer program that can run on the processor 520. When the processor 520 executes the computer program, it implements the method described in the above embodiments. The number of memories 510 and processors 520 can be one or more.

[0167] The electronic device also includes:

[0168] The communication interface 530 is used to communicate with external devices and exchange and transmit data.

[0169] If the memory 510, processor 520, and communication interface 530 are implemented independently, they can be interconnected via a bus to communicate with each other. This bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. This bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 5 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0170] Optionally, in a specific implementation, if the memory 510, processor 520, and communication interface 530 are integrated on a single chip, then the memory 510, processor 520, and communication interface 530 can communicate with each other through an internal interface.

[0171] This disclosure provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the methods provided in this disclosure.

[0172] This disclosure also provides a chip, which includes a processor for calling and executing instructions stored in a memory, causing a communication device on which the chip is installed to perform the methods provided in this disclosure.

[0173] This disclosure also provides a chip, including: an input interface, an output interface, a processor, and a memory. The input interface, output interface, processor, and memory are connected through an internal connection path. The processor is used to execute code in the memory. When the code is executed, the processor is used to execute the method provided in the application embodiment.

[0174] It should be understood that the aforementioned processor can be a Central Processing Unit (CPU), or other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. General-purpose processors can be microprocessors or any conventional processor. It is worth noting that the processor can be a processor supporting Advanced Reduced Instruction Set Machines (ARM) architecture.

[0175] Further, optionally, the aforementioned memory may include read-only memory and random access memory, and may also include non-volatile random access memory. The memory may be volatile or non-volatile, or may include both. Non-volatile memory may include read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory may include random access memory (RAM), which serves as an external cache. Many forms of RAM are available by way of example, but not limitation. Examples include Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), and Direct Rambus RAM (DR RAM).

[0176] In the above embodiments, implementation can be achieved, in whole or in part, by software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to this disclosure is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another.

[0177] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this disclosure. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0178] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this disclosure, "a plurality of" means two or more, unless otherwise explicitly specified.

[0179] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process. Furthermore, the scope of the preferred embodiments of this disclosure includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functionality involved.

[0180] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus or device (such as a computer-based system, a processor-included system or other system that can fetch and execute instructions from, an instruction execution system, apparatus or device).

[0181] It should be understood that various parts of this disclosure can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. All or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware, the program being stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiments.

[0182] Furthermore, the functional units in the various embodiments of this disclosure can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. This storage medium can be a read-only memory, a disk, or an optical disk, etc.

[0183] The above are merely specific embodiments of this disclosure, but the scope of protection of this disclosure is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in this disclosure, and these should all be included within the scope of protection of this disclosure. Therefore, the scope of protection of this disclosure should be determined by the scope of the claims.

Claims

1. A method for recommending navigation routes, characterized in that, include: Obtain historical travel data for candidate navigation routes, including road segments; Based on the historical travel data of the road segment and the route planning data of the candidate navigation routes, predictive data for the road segment is constructed. The predictive data includes the planning feature data of the road segment. The route planning data is the data that applications and service systems with map navigation functions need to use when planning candidate navigation routes. Based on the predicted data of the road segments included in the candidate navigation route, a recommended predicted value for the candidate navigation route is determined.

2. The method according to claim 1, characterized in that, The planning feature data of the road segment includes: time features and distance features. The step of constructing predictive data for the road segment based on historical travel data and route planning data of candidate navigation routes includes: Based on the historical travel time included in the historical travel data of the road segment and the route planning time included in the route planning data of the candidate navigation route, the time feature corresponding to the historical travel time is determined; Based on the historical origin and destination corresponding to the historical travel time in the historical travel data of the road segment, and the planned origin and destination included in the route planning data of the candidate navigation route, the distance feature corresponding to the historical travel time is determined; Predictive data for the road segment is constructed based at least on the time features and the distance features.

3. The method according to claim 2, characterized in that, The distance features include a first distance feature and a second distance feature. The distance features corresponding to the historical travel time are determined based on the historical origin and destination corresponding to the historical travel time in the historical travel data of the road segment, and the planned origin and destination included in the route planning data of the candidate navigation route. This includes: The distance between the historical starting point and the planned starting point is determined as the first distance feature; The distance between the historical endpoint and the planned endpoint is determined as the second distance feature.

4. The method according to claim 2, characterized in that, The planning feature data for the road segment also includes road condition features, and the method further includes: Based on the historical travel data of the road segment, the historical traffic information corresponding to the historical travel time, and the planned traffic information of the road segment included in the route planning data, the traffic characteristics corresponding to the historical travel time are determined.

5. The method according to claim 4, characterized in that, The planning feature data of the road segment also includes road segment proportion features, and the method further includes: The road segment proportion feature is determined based on the length proportion of the road segment in the candidate navigation routes.

6. The method according to claim 5, characterized in that, The predicted data for the road segment is constructed based at least on the time features and the distance features, including: The time characteristics, distance characteristics, road condition characteristics, and the proportion of the road segment corresponding to the same historical travel time are used to construct a prediction data for the road segment.

7. The method according to any one of claims 1-6, characterized in that, Based on the predicted data of the road segments included in the candidate navigation routes, the recommended predicted value of the candidate navigation routes is determined, including: The predicted data of the candidate navigation route, including road segments, is input into the trained familiar route prediction model to obtain the model prediction value corresponding to the predicted data of the road segments. Based on the model prediction values ​​corresponding to the prediction data of the road segment, determine the road segment prediction value; Based on the predicted values ​​of the road segments included in the candidate navigation route, a recommended predicted value for the candidate navigation route is determined.

8. The method according to claim 7, characterized in that, Based on the model prediction values ​​corresponding to the prediction data of the road segment, determine the predicted value of the road segment, including: The predicted values ​​of the model corresponding to the predicted data of the road segment are weighted and calculated to obtain the predicted value of the road segment. Based on the predicted values ​​of the road segments included in the candidate navigation route, a recommended predicted value for the candidate navigation route is determined, including: The recommended prediction value of the candidate navigation route is obtained by weighting the predicted values ​​of the road segments included in the candidate navigation route.

9. A training method for a familiar route prediction model, characterized in that, include: Obtain sample navigation routes, including historical travel data for sample road segments; Based on the sample historical travel data of the sample road segment and the sample route planning data of the sample navigation route, a training sample is constructed for the sample road segment. The training sample includes the sample planning feature data of the sample road segment. The route planning data is the data that applications and service systems with map navigation functions need to use when planning candidate navigation routes. The familiar route prediction model is trained using the training samples of the sample road segments. The familiar route prediction model is used to determine the recommended prediction value of the candidate navigation route based on the prediction data of the road segments included in the candidate navigation route.

10. The method according to claim 9, characterized in that, The familiar route prediction model is trained using training samples from the aforementioned sample road segments, including: The training samples of the sample road segment are input into the familiar road prediction model to obtain the model prediction value corresponding to the training samples of the sample road segment. Based on the model prediction values ​​corresponding to the training samples of the sample road segments, the sample navigation routes, and the sample travel routes, the parameters of the familiar route prediction model are adjusted.

11. The method according to claim 10, characterized in that, Based on the model prediction values ​​corresponding to the training samples of the sample road segments, the sample navigation routes, and the sample travel routes, the parameters of the familiar route prediction model are adjusted, including: Determine the travel coverage rate of the sample navigation route, where the travel coverage rate is the percentage of the length of the overlapping segments of the sample navigation route and the sample travel route within the sample navigation route; Based on the model prediction values ​​corresponding to the training samples of the sample road segments, determine the sample recommendation prediction values ​​for the sample navigation routes; The loss value of the familiar route prediction model is determined based on the sample recommendation prediction value of the sample navigation route and the travel coverage of the sample navigation route. Based on the loss value, the parameters of the familiar route prediction model are adjusted.

12. A navigation route recommendation device, characterized in that, include: The first acquisition module is used to acquire historical travel data of candidate navigation routes, including road segments. The first construction module is used to construct prediction data for the road segment based on the historical travel data of the road segment and the route planning data of the candidate navigation routes. The prediction data includes the planning feature data of the road segment. The route planning data is the data that the application and its service system with map navigation function need to use when planning candidate navigation routes. The determination module is used to determine the recommended predicted value of the candidate navigation route based on the predicted data of the road segments included in the candidate navigation route.

13. A training device for a familiar route prediction model, characterized in that, include: The second acquisition module is used to acquire sample navigation routes, including sample road segments, and sample historical travel data. The second construction module is used to construct training samples for the sample road segments based on the sample historical travel data of the sample road segments and the sample route planning data of the sample navigation routes. The training samples include sample planning feature data of the sample road segments. The route planning data is the data that applications and service systems with map navigation functions need to use when planning candidate navigation routes. The training module is used to train the familiar route prediction model using training samples from the sample road segments. The familiar route prediction model is used to determine the recommended prediction value of the candidate navigation route based on the prediction data of the road segments included in the candidate navigation route.

14. A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method of any one of claims 1-11.

Images