Yaw planning method and device, electronic equipment, storage medium and program product

Abstract

The embodiments of the present disclosure disclose a yaw planning method and device, electronic equipment, storage medium and program product. The method comprises: in response to detecting that a user deviates, determining a yaw location and an original navigation route endpoint; determining a plurality of first candidate yaw planning routes with the yaw location as the starting point and the original navigation route endpoint as the endpoint; determining a first historical behavior feature of the user and a first route feature of the plurality of first candidate yaw planning routes, inputting the first historical behavior feature and the first route feature into a pre-trained target yaw planning model to obtain a plurality of second candidate yaw planning routes; performing a preset filtering process on the plurality of second candidate yaw planning routes to obtain a first preset number of target yaw planning routes; and performing yaw planning on the user according to the target yaw planning routes. The technical solution can take into account the multiple intentions of taking a familiar route and returning to the original route when the user deviates, which is beneficial to improving the user experience and the service quality of the navigation platform.

Description

Technical Field

[0001] This disclosure relates to the field of traffic data processing technology, specifically to a yaw planning method, apparatus, electronic device, storage medium, and program product. Background Technology

[0002] With the development and advancement of society and navigation technology, many users rely on navigation route guidance when traveling. In daily use, deviations from the original route frequently occur due to users' habit of following familiar routes or momentary oversights leading to the wrong route. Current technologies typically try to guide users back to their original path when they deviate, but at this point, users may want to continue along their familiar routes and not return to their original path. Therefore, the current single deviation guidance solution is not conducive to improving the user experience and the service quality of the navigation platform. Summary of the Invention

[0003] This disclosure provides a yaw planning method, apparatus, electronic device, storage medium, and program product.

[0004] Firstly, this disclosure provides a yaw planning method.

[0005] Specifically, the yaw planning method includes:

[0006] In response to detecting a user deviation, determine the deviation location and the endpoint of the original navigation route;

[0007] Starting from the deviance point and ending at the end of the original navigation route, multiple first candidate deviance planning routes are determined;

[0008] Determine the user's first historical behavior characteristics and the first route characteristics of the multiple first candidate yaw planning routes. Input the user's first historical behavior characteristics, the multiple first candidate yaw planning routes and their first route characteristics into the pre-trained target yaw planning model to obtain multiple second candidate yaw planning routes.

[0009] The multiple second candidate yaw planning routes are subjected to a preset filtering process to obtain a first preset number of target yaw planning routes.

[0010] The user performs yaw planning based on the target yaw planning route.

[0011] In one implementation of this disclosure, after determining multiple first candidate yaw planning routes, the method further includes:

[0012] When the number of the first candidate yaw planning routes is greater than a preset number threshold, the user's second historical behavior characteristics are determined, and the user's second historical behavior characteristics, the first candidate yaw planning routes and their first route characteristics are input into the pre-trained target route evaluation model to obtain the second preset number of third candidate yaw planning routes with the highest second evaluation score.

[0013] Determine the second route characteristics of the third candidate yaw planning route, replace the first candidate yaw planning route with the third candidate yaw planning route, and replace the first route characteristics with the second route characteristics.

[0014] One implementation of this disclosure also includes:

[0015] The target yaw planning model is obtained through training;

[0016] The training to obtain the target yaw planning model includes:

[0017] Determine the initial yaw planning model;

[0018] A yaw planning training dataset is generated, comprising: a user's first training behavior feature, a user's first historical planned route, a third route feature corresponding to the user's first historical planned route, a user's first historical actual route, and a first evaluation score for the user's first historical planned route. The first evaluation score is an evaluation score obtained based on the user's actual route coverage and the actual route coverage of the original main route. The user's actual route coverage is the ratio between the distance of the overlapping part of the planned route and the user's first historical actual route and the distance actually traveled by the user during the current planning. The actual route coverage is the ratio between the distance of the yaw planning route and the overlapping part of the yaw route from the yaw point to the end point of the original main route and the distance of the yaw route.

[0019] Using the user's first training behavior features, the user's first historical planned route, the third route features, and the user's first historical actual route as inputs, the first evaluation score of the user's first historical planned route as output, and the objective function being to maximize the combination of the user's actual coverage rate and the original main route's actual coverage rate, the initial yaw planning model is trained to obtain the target yaw planning model.

[0020] In one implementation of this disclosure, the combined value of the user's actual coverage rate and the original main route's actual coverage rate is the sum of the product of the user's actual coverage rate and a first preset coefficient, and the product of the original main route's actual coverage rate and a second preset coefficient, wherein the sum of the first preset coefficient and the second preset coefficient is 1.

[0021] One implementation of this disclosure also includes:

[0022] When the distance of the eccentric route is less than or equal to the first preset distance, the actual coverage rate of the original main route is set to 0.

[0023] When the distance of the eccentric route is greater than the first preset distance but less than the second preset distance, the eccentric point is set as a point on the eccentric route that has traveled a third preset distance along the eccentric route; when the distance of the eccentric route is greater than the second preset distance, the eccentric point is set as a point on the eccentric route that has traveled a preset ratio along the eccentric route; based on the adjustment of the eccentric point, the actual coverage rate of the original main route is recalculated.

[0024] One implementation of this disclosure also includes:

[0025] The target route evaluation model is obtained through training;

[0026] The training of the target route evaluation model includes:

[0027] Determine the initial route evaluation model;

[0028] A route evaluation training dataset is generated, wherein the route evaluation training dataset includes user second training behavior features, user second historical planned routes, fourth route features corresponding to user second historical planned routes, user second historical actual routes, and a second evaluation score of user second historical planned routes. The second evaluation score is an evaluation score obtained based on the fourth route features and user actual coverage rate, wherein the user actual coverage rate is the ratio between the distance of the overlapping part of the planned route and the user second historical actual route and the user's actual walking distance.

[0029] Using the user's second training behavior features, the user's second historical planned route, the fourth route features, and the user's second historical actual route as inputs, the second evaluation score of the user's second historical planned route as output, and the objective function being that the route with the highest actual coverage rate is located in the third preset number of route queues with the highest second evaluation scores, the initial route evaluation model is trained to obtain the target route evaluation model.

[0030] Secondly, this disclosure provides a yaw planning device.

[0031] Specifically, the yaw planning device includes:

[0032] The first determination module is configured to determine the deviance location and the endpoint of the original navigation route in response to detecting a user deviation.

[0033] The second determining module is configured to determine multiple first candidate deviation planning routes, starting from the deviation location and ending at the end of the original navigation route.

[0034] The route generation module is configured to determine the user's first historical behavior characteristics and the first route characteristics of the multiple first candidate yaw planning routes, and input the user's first historical behavior characteristics, the multiple first candidate yaw planning routes and their first route characteristics into a pre-trained target yaw planning model to obtain multiple second candidate yaw planning routes.

[0035] The filtering module is configured to perform preset filtering on the multiple second candidate yaw planning routes to obtain a first preset number of target yaw planning routes.

[0036] The planning module is configured to perform yaw planning for the user based on the target yaw planning route.

[0037] Thirdly, embodiments of this disclosure provide an electronic device including a memory and at least one processor, wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the at least one processor to implement the method steps of the yaw planning method described above.

[0038] Fourthly, embodiments of this disclosure provide a computer-readable storage medium for storing computer instructions used by a yaw planning device, including computer instructions for performing the aforementioned yaw planning method in connection with the yaw planning device.

[0039] Fifthly, embodiments of this disclosure provide a computer program product, including a computer program / instructions, wherein the computer program / instructions, when executed by a processor, implement the method steps of the yaw planning method described above.

[0040] The technical solutions provided in this disclosure may have the following beneficial effects:

[0041] The above technical solution provides a yaw planning method. This method utilizes a yaw planning model trained by comprehensively considering the user's actual travel coverage and the actual travel coverage of the original main route to plan the yaw route. When a user deviates from the route, this technical solution can balance the user's desire to follow a familiar route and the desire to return to the original route, which helps improve the user experience and the service quality of the navigation platform.

[0042] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0043] Other features, objects, and advantages of this disclosure will become more apparent from the following detailed description of non-limiting embodiments, taken in conjunction with the accompanying drawings. In the drawings:

[0044] Figure 1 A flowchart illustrating a yaw planning method according to an embodiment of the present disclosure is shown;

[0045] Figure 2 A structural block diagram of a yaw planning device according to an embodiment of the present disclosure is shown;

[0046] Figure 3 A structural block diagram of an electronic device according to an embodiment of the present disclosure is shown;

[0047] Figure 4 This is a schematic diagram of the structure of a computer system suitable for implementing the yaw planning method according to an embodiment of the present disclosure. Detailed Implementation

[0048] In the following, exemplary embodiments of the present disclosure will be described in detail with reference to the accompanying drawings to enable those skilled in the art to readily implement them. Furthermore, for clarity, portions unrelated to the description of the exemplary embodiments have been omitted from the drawings.

[0049] In this disclosure, it should be understood that terms such as “comprising” or “having” are intended to indicate the presence of features, figures, steps, behaviors, components, parts or combinations thereof disclosed in this specification, and are not intended to exclude the possibility of the presence or addition of one or more other features, figures, steps, behaviors, components, parts or combinations thereof.

[0050] It should also be noted that, unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other. This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0051] The technical solution provided in this disclosure offers a yaw planning method. This method utilizes a yaw planning model trained by comprehensively considering the user's actual travel coverage and the actual travel coverage of the original main route to plan the yaw route. When a user deviates from the intended route, this technical solution can accommodate both the user's familiar route and the desire to return to the original route, thereby improving the user experience and the service quality of the navigation platform.

[0052] Figure 1 A flowchart illustrating a yaw planning method according to an embodiment of the present disclosure is shown, as follows: Figure 3 As shown, the yaw planning method includes the following steps S101-S105:

[0053] In step S101, in response to detecting user deviation, the deviation location and the endpoint of the original navigation route are determined;

[0054] In step S102, multiple first candidate deviation planning routes are determined, starting from the deviation location and ending at the original navigation route.

[0055] In step S103, the user's first historical behavior feature and the first route feature of the multiple first candidate yaw planning routes are determined. The user's first historical behavior feature, the multiple first candidate yaw planning routes and their first route features are input into the pre-trained target yaw planning model to obtain multiple second candidate yaw planning routes.

[0056] In step S104, the multiple second candidate yaw planning routes are subjected to preset filtering to obtain a second preset number of target yaw planning routes.

[0057] In step S105, yaw planning is performed on the user according to the target yaw planning route.

[0058] As mentioned above, with the development and advancement of society and navigation technology, many users rely on navigation route guidance when traveling. In daily use, deviations from the original route frequently occur due to users' habit of following familiar routes or momentary oversights leading to the wrong route. Current technologies typically try to guide users back to their original path when they deviate, but at this point, users may want to follow their familiar routes and not return to their original path. Therefore, the current single deviation guidance solution is not conducive to improving the user experience and the service quality of the navigation platform.

[0059] To address the aforementioned shortcomings, this implementation proposes a yaw planning method. This method utilizes a yaw planning model trained by comprehensively considering both the user's actual route coverage and the original main route coverage to plan the yaw route. When a user deviates from the intended route, this solution can accommodate both the user's familiar route and the desire to return to the original route, thereby improving the user experience and the service quality of the navigation platform.

[0060] In one embodiment of this disclosure, the yaw planning method can be applied to computers, computing devices, electronic devices, servers, service clusters, etc., for planning yaw routes.

[0061] In one embodiment of this disclosure, the detection of user deviation can be determined based on the distance between the user's actual location and a location on the planned route. For example, if the distance between the user's actual location and a location on the planned route exceeds a fourth preset distance, it can be determined that the user has deviated from the planned route. The location of the deviation can be determined as the location on the original main route at this time, that is, the location on the original main route when the user's deviation is detected.

[0062] In one embodiment of this disclosure, the first candidate yaw planning route refers to an optional yaw planning route generated from the yaw location as the starting point and the end point of the original navigation route as the ending point, which can be used to guide the user. There can be multiple first candidate yaw planning routes, which can be subsequently selected based on the target yaw planning model.

[0063] In one embodiment of this disclosure, the user's first historical behavior feature refers to a feature used to characterize the user's historical behavior. This first historical behavior feature may include one or more of the following features: the proportion of the user's historical travel using highways, the proportion of the user's historical travel using planned routes, the proportion of the user's historical travel using familiar routes, the overlap between the user's historically used familiar routes and corresponding planned routes, the user's historical starting and ending points, etc.

[0064] In one embodiment of this disclosure, the first route feature of the first candidate yaw planning route refers to the route feature extracted from the first candidate yaw planning route. The first route feature may include one or more of the following features: the number of traffic lights in the route, the estimated arrival time of the route, the proportion of highway sections in the route, the proportion of main roads in the route, the number of turns in the route, etc.

[0065] In one embodiment of this disclosure, the target yaw planning model refers to a pre-trained model used to generate yaw planning routes.

[0066] In one embodiment of this disclosure, the preset filtering process refers to filtering multiple second candidate evasion planning routes to obtain a first preset number of routes most suitable for the user. The preset filtering process may include one or more of the following filtering methods: such as route similarity, the proportion of highway sections in the route, the proportion of main roads in the route, the proportion of side roads in the route, the number of traffic lights in the route, the number of turns in the route, etc. For example, only one second candidate evasion planning route with high similarity may be retained; second candidate evasion planning routes with too low a proportion of highway sections, too low a proportion of main roads, too high a proportion of side roads, too many traffic lights, or too many turns may be deleted, etc. The first preset number can be set according to the needs of actual application, and this disclosure does not impose any particular limitation on it; for example, the first preset number may be set to 3.

[0067] In the above implementation, when performing yaw planning for a user, after detecting a user's yaw, the yaw location and the endpoint of the original navigation route are first determined; then, using the yaw location as the starting point and the endpoint of the original navigation route as the endpoint, multiple first candidate yaw planning routes are recalled; then, the user's first historical behavior characteristics and the first route characteristics of the multiple first candidate yaw planning routes are determined, and the user's first historical behavior characteristics, the multiple first candidate yaw planning routes and their first route characteristics are input into a pre-trained target yaw planning model to obtain multiple second candidate yaw planning routes; then, the multiple second candidate yaw planning routes are subjected to a preset filtering process to obtain a first preset number of target yaw planning routes; finally, the target yaw planning routes are provided to the user for selection, and yaw planning and guidance are performed on the user based on the yaw planning route finally selected by the user.

[0068] In one embodiment of this disclosure, after determining multiple first candidate yaw planning routes in step S102, the following steps may also be included:

[0069] When the number of the first candidate yaw planning routes is greater than a preset number threshold, the user's second historical behavior characteristics are determined, and the user's second historical behavior characteristics, the first candidate yaw planning routes and their first route characteristics are input into the pre-trained target route evaluation model to obtain the second preset number of third candidate yaw planning routes with the highest second evaluation score.

[0070] Determine the second route characteristics of the third candidate yaw planning route, replace the first candidate yaw planning route with the third candidate yaw planning route, and replace the first route characteristics with the second route characteristics.

[0071] In one embodiment of this disclosure, the target route evaluation model refers to a pre-trained model used to evaluate routes.

[0072] In this implementation, when the number of identified first candidate yaw planning routes is large, for example, exceeding a preset threshold, a trained target route evaluation model can be used to further filter the large number of first candidate yaw planning routes to optimize the data foundation for the target yaw planning routes subsequently provided to the user. Specifically, the user's second historical behavior characteristics are determined, and the user's second historical behavior characteristics, the first candidate yaw planning routes, and their first route characteristics are input into the pre-trained target route evaluation model to obtain the second preset number of third candidate yaw planning routes with the highest second evaluation scores.

[0073] The user's second historical behavior characteristics may include one or more of the following characteristics: the proportion of the user's historical trips using highways, the proportion of the user's historical trips using planned routes, the proportion of the user's historical trips using familiar routes, the overlap between the familiar routes used by the user in the past and the corresponding planned routes, the starting points and ending points used by the user in the past, etc.

[0074] The preset quantity threshold and the second preset quantity can be set according to the needs of actual application. This disclosure does not impose any special limitations on them. For example, the preset quantity threshold and the second preset quantity can both be set to 20.

[0075] In one embodiment of this disclosure, the second route feature of the third candidate yaw planning route refers to the route feature extracted from the third candidate yaw planning route. The second route feature may include one or more of the following features: the number of traffic lights in the route, the estimated arrival time of the route, the proportion of highway sections in the route, the proportion of main roads in the route, the number of turns in the route, etc.

[0076] In this embodiment, the first candidate yaw planning route is replaced by the third candidate yaw planning route, and the first route feature is replaced by the second route feature and input into the target yaw planning model. That is, the determined user's first historical behavior feature, the third candidate yaw planning route and its second route feature are input into the pre-trained target yaw planning model to obtain multiple second candidate yaw planning routes.

[0077] In one embodiment of this disclosure, the method may further include the following steps:

[0078] The target yaw planning model is obtained through training;

[0079] The step of training the target yaw planning model may include the following steps:

[0080] Determine the initial yaw planning model;

[0081] A yaw planning training dataset is generated, comprising: a user's first training behavior feature, a user's first historical planned route, a third route feature corresponding to the user's first historical planned route, a user's first historical actual route, and a first evaluation score for the user's first historical planned route. The first evaluation score is an evaluation score obtained based on the user's actual route coverage and the actual route coverage of the original main route. The user's actual route coverage is the ratio between the distance of the overlapping part of the planned route and the user's first historical actual route and the distance actually traveled by the user during the current planning. The actual route coverage is the ratio between the distance of the yaw planning route and the overlapping part of the yaw route from the yaw point to the end point of the original main route and the distance of the yaw route.

[0082] Using the user's first training behavior features, the user's first historical planned route, the third route features, and the user's first historical actual route as inputs, the first evaluation score of the user's first historical planned route as output, and the objective function being to maximize the combination of the user's actual coverage rate and the original main route's actual coverage rate, the initial yaw planning model is trained to obtain the target yaw planning model.

[0083] In one embodiment of this disclosure, the initial yaw planning model can be selected according to the needs of the actual application. For example, the initial yaw planning model can be selected as a deep learning model such as DNN (Deep Neural Network), or other types of models, etc.

[0084] In one embodiment of this disclosure, the first user training behavior feature refers to a feature used as input data for training the yaw planning model, which characterizes the user's behavioral characteristics. This first user training behavior feature may include one or more of the following features: the proportion of the user using highways during travel, the proportion of the user using planned routes during travel, the proportion of the user using familiar routes during travel, the overlap between familiar routes used by the user and corresponding planned routes, the user's previously used starting and ending points, etc.

[0085] In one embodiment of this disclosure, the user's first historical planned route refers to a planned route generated within a preset historical time period based on the user-set start and end points, which serves as input data for training the yaw planning model.

[0086] In one embodiment of this disclosure, the third route feature corresponding to the user's first historical planned route refers to the route feature extracted from the user's first historical planned route, which serves as the input data for training the yaw planning model. The third route feature may include one or more of the following features: the number of traffic lights in the route, the estimated arrival time of the route, the proportion of highways in the route, the proportion of main roads in the route, the number of turns in the route, etc.

[0087] In one embodiment of this disclosure, the user's first historical actual route refers to the route actually traveled by the user between the starting point and the destination set by the user within a preset historical time period, which serves as the input data for training the yaw planning model.

[0088] In one embodiment of this disclosure, the first evaluation score of the user's first historical planned route refers to the evaluation score obtained by evaluating the user's first historical planned route based on two factors: the user's actual coverage rate and the actual coverage rate of the original main route, which are used as the output data of the yaw planning model training.

[0089] In one embodiment of this disclosure, the user's actual travel coverage rate refers to the ratio between the distance of the overlapping portion of the planned route and the user's first historical traveled route during a route planning session, and the user's actual travel distance during the current planning session. The user's actual travel distance refers to the distance actually traveled by the user at the end of the current navigation session.

[0090] In one embodiment of this disclosure, the actual coverage rate of the original main route is the ratio between the distance of the eccentric planned route and the overlapping portion of the eccentric route from the eccentric point to the end point of the original main route, and the distance of the eccentric route. The eccentric planned route refers to the route replanned after detecting a user eccentricity. The original main route refers to the main route planned before the user eccentricated. The eccentric point refers to the location on the original main route corresponding to the user's eccentricity during the guidance process of the original main route, and the eccentric route refers to the route from the eccentric point to the end point of the original main route.

[0091] In the above embodiments, when training the yaw planning model, an initial yaw planning model is first determined; then, a yaw planning training dataset is generated, consisting of the user's first training behavior features, the user's first historical planned route, the third route features corresponding to the user's first historical planned route, the user's first historical actual route, and the first evaluation score of the user's first historical planned route; then, using the user's first training behavior features, the user's first historical planned route, the third route features, and the user's first historical actual route as input, and the first evaluation score of the user's first historical planned route as output, with the objective function being to maximize the combination of the user's actual coverage rate and the original main route's actual coverage rate, the initial yaw planning model is trained. When the training results converge, the trained target yaw planning model is obtained. The learning and training of the above yaw planning model can be implemented using learning and training methods mastered by those skilled in the art, and this disclosure does not particularly limit the specific learning and training implementation method of yaw planning.

[0092] As mentioned above, the objective function when training the yaw planning model is to maximize the combined value of the user's actual coverage rate and the original main route's actual coverage rate. This combined value is the sum of the product of the user's actual coverage rate and a first preset coefficient, and the product of the original main route's actual coverage rate and a second preset coefficient. The sum of the first and second preset coefficients is 1. Therefore, the combined value of the user's actual coverage rate and the original main route's actual coverage rate can be expressed as:

[0093] α1·coverage+α2·mainpathcoverage,

[0094] Wherein, α1 is the first preset coefficient, with a value range of 0-1, α2 is the second preset coefficient, with a value range of 0-1, and α1+α2=1, coverage is the actual coverage rate of the user, and mainpathcoverage is the actual coverage rate of the original main route.

[0095] Considering that if the deviation route itself is relatively short, it is not suitable to adjust the user's route in this case, so as not to impair the user's experience. Therefore, in one embodiment of this disclosure, when the distance of the deviation route is less than or equal to a first preset distance, the actual coverage rate of the original main route is set to 0, that is, the user's own driving intention is mainly considered. The first preset distance can be set according to the needs of actual application. For example, the first preset distance can be set to 5 kilometers.

[0096] Meanwhile, considering that the user's driving intention should be valued in the early stages of deviation, it is not suitable to adjust the user's route at this time to avoid damaging the user experience. If adjustments are necessary, they should be made later. Therefore, in one embodiment of this disclosure, when the distance of the deviation route is greater than a first preset distance but less than a second preset distance, the deviation point can be reset to a point on the deviation route where the user has traveled a third preset distance along the deviation route. Subsequently, the actual coverage rate of the original main route is recalculated based on the new deviation point. The second and third preset distances can be set according to the actual application needs. For example, the second preset distance can be set to 40 kilometers, and the third preset distance can be set to the same as the first preset distance, i.e., 5 kilometers. For example, if the distance of the deviation route is 10 kilometers, the deviation point can be reset to 5 kilometers away from the starting point of the deviation route, i.e., 5 kilometers from the original deviation point. When the distance of the eccentric route is greater than the second preset distance, the eccentric point can be reset to a point located at a preset percentage of travel along the eccentric route. The preset percentage can be set according to the needs of the actual application; for example, it can be set to 10%. For instance, if the distance of the eccentric route is 100 kilometers, the eccentric point can be reset to 10% of the distance from the starting point of the eccentric route, i.e., 10 kilometers from the original eccentric point. Subsequently, the actual coverage rate of the original main route will be recalculated based on the new eccentric point.

[0097] By using the objective function that maximizes the combination of user actual coverage rate and original main route actual coverage rate, and by adjusting the deviation point, it is possible to achieve the following in deviation scenarios: the first half of the deviation route tends to be the user's familiar route mode, and the second half of the deviation route tends to be the user's return to the original route mode. This can both ensure that the user's wishes are reflected and reduce negative feedback from users who change routes after deviation.

[0098] In one embodiment of this disclosure, the method may further include the following steps:

[0099] The target route evaluation model is obtained through training;

[0100] The step of training the target route evaluation model may include the following steps:

[0101] Determine the initial route evaluation model;

[0102] A route evaluation training dataset is generated, wherein the route evaluation training dataset includes user second training behavior features, user second historical planned routes, fourth route features corresponding to user second historical planned routes, user second historical actual routes, and a second evaluation score of user second historical planned routes. The second evaluation score is an evaluation score obtained based on the fourth route features and user actual coverage rate, wherein the user actual coverage rate is the ratio between the distance of the overlapping part of the planned route and the user second historical actual route and the user's actual walking distance.

[0103] Using the user's second training behavior features, the user's second historical planned route, the fourth route features, and the user's second historical actual route as inputs, the second evaluation score of the user's second historical planned route as output, and the objective function being that the route with the highest actual coverage rate is located in the third preset number of route queues with the highest second evaluation scores, the initial route evaluation model is trained to obtain the target route evaluation model.

[0104] In one embodiment of this disclosure, the initial route evaluation model can be selected according to the needs of the actual application. For example, the initial route evaluation model can be selected as a deep learning model such as DNN (Deep Neural Network), or other types of models, etc.

[0105] In one embodiment of this disclosure, the second user training behavior feature is similar to the first user training behavior feature, referring to features used as input data for training the route evaluation model to characterize user behavior. The second user training behavior feature may include one or more of the following features: the proportion of users using highways when traveling, the proportion of users using planned routes when traveling, the proportion of users using familiar routes when traveling, the overlap between familiar routes used by users and corresponding planned routes, the starting and ending points used by users, etc.

[0106] In one embodiment of this disclosure, the user's second historical planned route is similar to the user's second training behavior feature, referring to the planned route generated within a preset historical time period based on the user's set start and end points, which serves as the input data for training the route evaluation model.

[0107] In one embodiment of this disclosure, the fourth route feature corresponding to the user's second historical planned route refers to the route feature extracted from the user's second historical planned route, which serves as input data for training the route evaluation model. The fourth route feature may include one or more of the following features: the number of traffic lights in the route, the estimated arrival time of the route, the proportion of highways in the route, the proportion of main roads in the route, the number of turns in the route, etc.

[0108] In one embodiment of this disclosure, the user's second historical actual route refers to the route actually traveled by the user between the starting point and the destination set by the user within a preset historical time period, which serves as input data for training the route evaluation model.

[0109] In one embodiment of this disclosure, the second evaluation score of the user's second historical planned route refers to the evaluation score obtained by evaluating the user's second historical planned route based on two factors: the fourth route feature and the user's actual walking coverage rate, which are used as output data for training the route evaluation model. During the evaluation, for example, a corresponding weight can be assigned to each feature in the fourth route feature and the user's actual walking coverage rate. Then, the user's second historical planned route is evaluated separately according to each feature in the fourth route feature and the user's actual walking coverage rate to obtain corresponding sub-evaluation values. Finally, the weighted sum of the sub-evaluation values ​​and their corresponding weights is calculated to obtain the evaluation score of the user's second historical planned route.

[0110] As mentioned above, the user's actual travel coverage rate refers to the ratio between the distance of the overlapping portion of the planned route and the user's first historical traveled route during a certain route planning, and the user's actual travel distance during the current planning. The user's actual travel distance refers to the distance actually traveled by the user at the end of this navigation session.

[0111] In the above embodiments, when training the route evaluation model, an initial route evaluation model is first determined; then, a route evaluation training dataset is generated, consisting of the user's second training behavior features, the user's second historical planned routes, the fourth route features corresponding to the user's second historical planned routes, the user's second historical actually traveled routes, and the second evaluation scores of the user's second historical planned routes; then, using the user's second training behavior features, the user's second historical planned routes, the fourth route features, and the user's second historical actually traveled routes as inputs, and the second evaluation scores of the user's second historical planned routes as outputs, with the objective function being that the route with the highest actual travel coverage is located in a third preset number of route queues with the highest second evaluation scores, the initial route evaluation model is trained. When the training results converge, the trained target route evaluation model is obtained. The learning and training of the above route evaluation model can be implemented using learning and training methods mastered by those skilled in the art, and this disclosure does not particularly limit the specific learning and training implementation method of route evaluation.

[0112] As mentioned above, the objective function for training the route evaluation model is to ensure that the route with the highest user coverage is among the top three preset number of routes with the highest second evaluation scores. Specifically, after sorting by second evaluation scores from highest to lowest, the route with the highest user coverage is among the top three preset number of routes. The third preset number can be set according to the needs of the actual application, and this disclosure does not impose any particular limitation on its specific value. For example, the third preset number can be set to 20, meaning the objective function can be described as the route with the highest user coverage being among the top 20 routes with the highest second evaluation scores.

[0113] The following are embodiments of the apparatus disclosed herein, which can be used to execute embodiments of the method disclosed herein.

[0114] Figure 2 The diagram shows a structural block diagram of a yaw planning device according to an embodiment of the present disclosure. This device can be implemented as part or all of an electronic device through software, hardware, or a combination of both. Figure 2 As shown, the yaw planning device includes:

[0115] The first determining module 201 is configured to determine the deviance location and the endpoint of the original navigation route in response to detecting a user deviation.

[0116] The second determining module 202 is configured to determine multiple first candidate deviation planning routes, starting from the deviation location and ending at the end of the original navigation route.

[0117] The route generation module 203 is configured to determine the user's first historical behavior characteristics and the first route characteristics of the multiple first candidate yaw planning routes, and input the user's first historical behavior characteristics, the multiple first candidate yaw planning routes and their first route characteristics into the pre-trained target yaw planning model to obtain multiple second candidate yaw planning routes.

[0118] The filtering module 204 is configured to perform preset filtering on the plurality of second candidate yaw planning routes to obtain a first preset number of target yaw planning routes.

[0119] The planning module 205 is configured to perform yaw planning for the user based on the target yaw planning route.

[0120] As mentioned above, with the development and advancement of society and navigation technology, many users rely on navigation route guidance when traveling. In daily use, deviations from the original route frequently occur due to users' habit of following familiar routes or momentary oversights leading to the wrong route. Current technologies typically try to guide users back to their original path when they deviate, but at this point, users may want to follow their familiar routes and not return to their original path. Therefore, the current single deviation guidance solution is not conducive to improving the user experience and the service quality of the navigation platform.

[0121] To address the aforementioned shortcomings, this embodiment proposes a yaw planning device. This device utilizes a yaw planning model trained by comprehensively considering the user's actual route coverage and the actual route coverage of the original main route to plan the yaw route. When a user deviates from the intended path, this technical solution can accommodate both the user's familiar route and the desire to return to the original route, thereby improving the user experience and the service quality of the navigation platform.

[0122] In one embodiment of this disclosure, the yaw planning device can be implemented as a computer, computing device, electronic device, server, service cluster, etc., for planning yaw routes.

[0123] The technical terms and features involved in the above-mentioned device embodiments are the same as or similar to those mentioned in the above-mentioned method embodiments. For explanations and descriptions of the technical terms and features involved in the above-mentioned device embodiments, please refer to the explanations of the above-mentioned method embodiments. They will not be repeated here.

[0124] This disclosure also discloses an electronic device. Figure 3 This diagram illustrates a structural block diagram of an electronic device according to an embodiment of the present disclosure, such as... Figure 3 As shown, the electronic device 300 includes a memory 301 and a processor 302; wherein,

[0125] The memory 301 is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor 302 to implement the above method steps.

[0126] Figure 4 This is a schematic diagram of the structure of a computer system suitable for implementing the yaw planning method according to an embodiment of the present disclosure.

[0127] like Figure 4As shown, the computer system 400 includes a processing unit 401, which can execute various processes described above based on a program stored in a read-only memory (ROM) 402 or a program loaded from a storage section 408 into a random access memory (RAM) 403. The RAM 403 also stores various programs and data required for the operation of the computer system 400. The processing unit 401, ROM 402, and RAM 403 are interconnected via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0128] The following components are connected to I / O interface 405: an input section 406 including a keyboard, mouse, etc.; an output section 407 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN card, modem, etc. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to I / O interface 405 as needed. A removable medium 411, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 410 as needed so that computer programs read from it can be installed into storage section 408 as needed. The processing unit 401 can be implemented as a CPU, GPU, TPU, FPGA, NPU, etc.

[0129] In particular, according to embodiments of this disclosure, the methods described above can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program tangibly embodied on a readable medium thereof, the computer program containing program code for performing the methods. In such embodiments, the computer program can be downloaded and installed from a network via communication section 409, and / or installed from removable medium 411.

[0130] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0131] The units or modules described in the embodiments of this disclosure can be implemented in software or hardware. The described units or modules can also be located in a processor, and the names of these units or modules do not necessarily constitute a limitation on the unit or module itself.

[0132] In another aspect, this disclosure also provides a computer-readable storage medium, which may be a computer-readable storage medium included in the apparatus described in the above embodiments; or it may be a standalone computer-readable storage medium not assembled into a device. The computer-readable storage medium stores one or more programs that are used by one or more processors to perform the methods described in this disclosure.

[0133] The above description is merely a preferred embodiment of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features disclosed in this disclosure that have similar functions.

Claims

1. A yaw planning method, comprising: In response to detecting a user deviation, determine the deviation location and the endpoint of the original navigation route; Starting from the deviance point and ending at the end of the original navigation route, multiple first candidate deviance planning routes are determined; The user's first historical behavior characteristics and the first route characteristics of the multiple first candidate eccentricity planning routes are determined. The user's first historical behavior characteristics, the multiple first candidate eccentricity planning routes and their first route characteristics are input into a pre-trained target eccentricity planning model to obtain multiple second candidate eccentricity planning routes. The target eccentricity planning model takes the user's first training behavior characteristics, the user's first historical planning route, the third route characteristics corresponding to the user's first historical planning route and the user's first historical actual route as input, takes the first evaluation score of the user's first historical planning route as output, and takes the maximum combination value of the user's actual coverage rate and the original main route's actual coverage rate as the objective function, and is trained accordingly. The multiple second candidate yaw planning routes are subjected to a preset filtering process to obtain a first preset number of target yaw planning routes. The user performs yaw planning based on the target yaw planning route.

2. The method according to claim 1, further comprising, after determining multiple first candidate yaw planning routes: When the number of the first candidate yaw planning routes is greater than a preset number threshold, the user's second historical behavior characteristics are determined, and the user's second historical behavior characteristics, the first candidate yaw planning routes and their first route characteristics are input into the pre-trained target route evaluation model to obtain the second preset number of third candidate yaw planning routes with the highest second evaluation score. Determine the second route characteristics of the third candidate yaw planning route, replace the first candidate yaw planning route with the third candidate yaw planning route, and replace the first route characteristics with the second route characteristics.

3. The method according to claim 1 or 2, further comprising: The target yaw planning model is obtained through training; The training to obtain the target yaw planning model includes: Determine the initial yaw planning model; A yaw planning training dataset is generated, comprising: a user's first training behavior feature, a user's first historical planned route, a third route feature corresponding to the user's first historical planned route, a user's first historical actual route, and a first evaluation score for the user's first historical planned route. The first evaluation score is an evaluation score obtained based on the user's actual route coverage and the actual route coverage of the original main route. The user's actual route coverage is the ratio between the distance of the overlapping part of the planned route and the user's first historical actual route and the distance actually traveled by the user during the current planning. The actual route coverage is the ratio between the distance of the yaw planning route and the overlapping part of the yaw route from the yaw point to the end point of the original main route and the distance of the yaw route. Using the user's first training behavior features, the user's first historical planned route, the third route features, and the user's first historical actual route as inputs, the first evaluation score of the user's first historical planned route as output, and the objective function being to maximize the combination of the user's actual coverage rate and the original main route's actual coverage rate, the initial yaw planning model is trained to obtain the target yaw planning model.

4. The method according to claim 3, wherein the combined value of the user's actual coverage rate and the original main route's actual coverage rate is the sum of the product of the user's actual coverage rate and a first preset coefficient, and the product of the original main route's actual coverage rate and a second preset coefficient, wherein, The sum of the first preset coefficient and the second preset coefficient is 1.

5. The method according to claim 4, further comprising: When the distance of the eccentric route is less than or equal to the first preset distance, the actual coverage rate of the original main route is set to 0. When the distance of the yaw route is greater than the first preset distance but less than the second preset distance, the yaw point is set as a point on the yaw route that has traveled a third preset distance along the yaw route direction; When the distance of the eccentric route is greater than the second preset distance, the eccentric point is set as a point on the eccentric route at a preset ratio of travel along the eccentric route direction; based on the adjustment of the eccentric point, the actual coverage rate of the original main route is recalculated.

6. The method according to claim 2, further comprising: The target route evaluation model is obtained through training; The training of the target route evaluation model includes: Determine the initial route evaluation model; A route evaluation training dataset is generated, wherein the route evaluation training dataset includes user second training behavior features, user second historical planned routes, fourth route features corresponding to user second historical planned routes, user second historical actual routes, and a second evaluation score of user second historical planned routes. The second evaluation score is an evaluation score obtained based on the fourth route features and user actual coverage rate, wherein the user actual coverage rate is the ratio between the distance of the overlapping part of the planned route and the user second historical actual route and the user's actual walking distance. Using the user's second training behavior features, the user's second historical planned route, the fourth route features, and the user's second historical actual route as inputs, the second evaluation score of the user's second historical planned route as output, and the objective function being that the route with the highest actual coverage rate is located in the third preset number of route queues with the highest second evaluation scores, the initial route evaluation model is trained to obtain the target route evaluation model.

7. A yaw planning device, comprising: The first determination module is configured to determine the deviance location and the endpoint of the original navigation route in response to detecting a user deviation. The second determining module is configured to determine multiple first candidate deviation planning routes, starting from the deviation location and ending at the end of the original navigation route. The route generation module is configured to determine the user's first historical behavior characteristics and the first route characteristics of the multiple first candidate eccentricity planning routes. It then inputs these characteristics into a pre-trained target eccentricity planning model to obtain multiple second candidate eccentricity planning routes. The target eccentricity planning model takes the user's first training behavior characteristics, the user's first historical planned route, the third route characteristics corresponding to the user's first historical planned route, and the user's first historical actual route as input, and the first evaluation score of the user's first historical planned route as output. The objective function is to maximize the combination of the user's actual coverage rate and the original main route's actual coverage rate. The filtering module is configured to perform preset filtering on the multiple second candidate yaw planning routes to obtain a first preset number of target yaw planning routes. The planning module is configured to perform yaw planning for the user based on the target yaw planning route.

8. An electronic device comprising a memory and at least one processor; wherein, The memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the at least one processor to implement the steps of the method according to any one of claims 1-6.

9. A computer-readable storage medium having computer instructions stored thereon, wherein, When executed by a processor, the computer instructions implement the steps of the method described in any one of claims 1-6.

10. A computer program product comprising a computer program / instructions, wherein, When the computer program / instructions are executed by the processor, they implement the steps of the method described in any one of claims 1-6.

Images