Route recommendation method and apparatus

Abstract

The specification discloses a route recommendation method and device. The method comprises: acquiring a plurality of reachable routes between start points, and determining an estimated arrival time of each reachable route; calculating a goodness rate of the estimated arrival time of each reachable route, the goodness rate representing a probability that an error of the estimated arrival time is within an acceptable range; for each reachable route, determining a time consumption score of the reachable route according to the estimated arrival time of the reachable route and the goodness rate of the estimated arrival time; and determining a reachable route with a time consumption score meeting a time consumption requirement as a recommended route between the start points. The above scheme can improve the accuracy of time consumption evaluation of the reachable routes, reduce the difference between the actual arrival time and the estimated arrival time of the recommended route as much as possible, thereby improving the accuracy of the recommended route and improving the user experience.

Description

Technical Field

[0001] This specification relates to the field of navigation technology, and in particular to a route recommendation method and apparatus. Background Technology

[0002] Electronic maps are increasingly used for route navigation in people's daily lives. For example, when a passenger takes a taxi, the driver can use an electronic map to plan the route and provide navigation. When planning a route, the estimated travel time for each route is usually estimated, and the route with the shortest travel time is recommended. However, in practice, the estimated travel time often differs from the actual travel time, resulting in poor accuracy in route recommendations and impacting the user experience. Summary of the Invention

[0003] In view of this, this specification provides a route recommendation method and apparatus.

[0004] Specifically, this specification is implemented through the following technical solution:

[0005] A route recommendation method includes:

[0006] Obtain several reachable routes between starting points and determine the estimated arrival time for each of the reachable routes;

[0007] Calculate the goodness rate of the estimated arrival time for each reachable route, where the goodness rate represents the probability that the error of the estimated arrival time is within an acceptable range;

[0008] For each reachable route, a time score is determined based on the estimated arrival time and the goodness rate of the estimated arrival time.

[0009] The reachable routes that meet the time consumption requirements are determined as the recommended routes between the starting points.

[0010] Optionally, the calculation of the goodness rate of the estimated arrival time for each reachable route includes:

[0011] Assuming that the actual arrival time of the reachable route follows a normal distribution with a mean of its expected arrival time, calculate the standard deviation of the normal distribution;

[0012] The normal distribution of the actual arrival time is standardized based on the estimated arrival time and the standard deviation, so that the actual arrival time follows a standard normal distribution.

[0013] Under the standard normal distribution, the good rate of the expected arrival time is calculated.

[0014] Optionally, the step of calculating the standard deviation of the normal distribution, assuming that the actual arrival time of the reachable route follows a normal distribution with a mean equal to its estimated arrival time, includes:

[0015] The route information of the reachable route is input into a trained time deviation prediction model, and the time deviation prediction model is used to predict the time deviation between the actual arrival time and the expected arrival time of the reachable route.

[0016] Assuming that the actual arrival time of the reachable route follows a normal distribution with a mean of its expected arrival time, the standard deviation of the normal distribution is calculated based on the time deviation.

[0017] Optionally, the training process of the time deviation prediction model includes:

[0018] Obtain sample routes, which have actual arrival time labels;

[0019] The route information of the sample route is input into the time deviation prediction model to be trained, and the time deviation prediction model is used to predict the time deviation between the actual arrival time and the expected arrival time of the sample route.

[0020] Calculate the true time deviation of the sample route based on the actual arrival time label and the estimated arrival time;

[0021] The time deviation prediction model is trained based on the difference between the time deviation predicted by the time deviation prediction model and the actual time deviation.

[0022] Optionally, calculating the good / good rate of the expected arrival time under the standard normal distribution includes:

[0023] The expected range of the actual arrival time is determined based on the favorable range of the estimated arrival time;

[0024] The probability value that the actual arrival time falls within the expected interval under the standard normal distribution is calculated as the good rate of the expected arrival time.

[0025] Optionally, the time requirement is to have the lowest time score, and determining the time score of the reachable route based on the estimated arrival time and the goodness rate of the estimated arrival time includes:

[0026] The non-excellent rate of the estimated arrival time is calculated based on the excellent rate of the estimated arrival time;

[0027] Determine a first weight for the estimated arrival time and a second weight for the non-excellent rate;

[0028] Based on the first weight and the second weight, the estimated arrival time and the non-excellent rate are weighted and summed to obtain the time score of the reachable route.

[0029] A navigation method for ride-hailing scenarios includes:

[0030] Get the specified starting point;

[0031] Based on the aforementioned route recommendation method, a target recommended route between the starting points is determined;

[0032] The target recommended route is recommended to the user so that the user can navigate based on the target recommended route.

[0033] A route recommendation device, comprising:

[0034] The route acquisition unit acquires several reachable routes between starting points and determines the estimated arrival time of each reachable route;

[0035] The goodness rate calculation unit calculates the goodness rate of the estimated arrival time for each reachable route, where the goodness rate represents the probability that the error of the estimated arrival time is within an acceptable range;

[0036] The time consumption determination unit determines the time consumption score of each reachable route based on the estimated arrival time and the goodness rate of the estimated arrival time.

[0037] The route recommendation unit determines the reachable routes that meet the time consumption requirements as recommended routes between the starting points.

[0038] An electronic device, comprising:

[0039] processor;

[0040] Memory used to store processor-executable instructions;

[0041] The processor implements the aforementioned method by running the executable instructions.

[0042] A computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the steps of the aforementioned method.

[0043] A computer program product, when executed by a processor, implements the aforementioned method.

[0044] Using the above implementation method, this specification can determine the estimated arrival time and the goodness rate of the estimated arrival time for each reachable route between starting points. Then, based on the estimated arrival time and the goodness rate, a time consumption score is determined for the corresponding reachable route. Finally, based on the time consumption score, a recommended route between the starting points is determined from the reachable routes. By introducing the goodness rate of the estimated arrival time, the accuracy of the time assessment for reachable routes can be improved, minimizing the difference between the actual arrival time and the estimated arrival time of the recommended route, thereby improving the accuracy of the recommended route and enhancing the user experience. Attached Figure Description

[0045] Figure 1 This is a flowchart illustrating an exemplary embodiment of a route recommendation method in this specification.

[0046] Figure 2 This is a flowchart illustrating an exemplary embodiment of the method for calculating the expected arrival time good rate.

[0047] Figure 3 This is a flowchart illustrating a navigation method in a ride-hailing scenario, as shown in an exemplary embodiment of this specification.

[0048] Figure 4 This is a hardware structure diagram of an electronic device containing a route recommendation device, as illustrated in an exemplary embodiment of this specification.

[0049] Figure 5 This is a block diagram illustrating an exemplary embodiment of a route recommendation device in this specification. Detailed Implementation

[0050] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this specification. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this specification as detailed in the appended claims.

[0051] The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of this specification. The singular forms “a,” “the,” and “the” as used in this specification and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0052] It should be understood that although the terms first, second, third, etc., may be used in this specification to describe various information, this information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this specification, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0053] Electronic maps are increasingly used for route navigation in people's daily lives. For example, when a passenger takes a taxi, the driver can use an electronic map to plan the route and provide navigation. When planning a route, the estimated travel time for each route is usually estimated, and the route with the shortest travel time is recommended. However, in practice, the estimated travel time often differs from the actual travel time, resulting in poor accuracy in route recommendations and impacting the user experience.

[0054] For example, suppose a user departs from Zhejiang University's Yuquan Campus and travels to Hangzhou Xiaoshan Airport. The recommended route is estimated to take 60 minutes, but it actually takes 90 minutes to reach Hangzhou Xiaoshan Airport, which is 30 minutes longer than the estimated time. This route recommendation method does not take into account the accuracy of the estimated time, which seriously affects the user experience.

[0055] This manual provides a route recommendation method that combines the estimated travel time and the accuracy rate of the estimated travel time to recommend routes to users, thereby improving the accuracy of route recommendations.

[0056] Figure 1 This is a flowchart illustrating an exemplary embodiment of a route recommendation method in this specification.

[0057] Please refer to Figure 1 The route recommendation method may include the following steps:

[0058] Step 102: Obtain several reachable routes between the starting points and determine the estimated arrival time of each reachable route.

[0059] In this specification, the starting point can be a user-specified origin and destination in a self-driving scenario. For example, if a user drives from Zhejiang University Yuquan Campus to Hangzhou Xiaoshan Airport to pick someone up, the starting point is Zhejiang University Yuquan Campus, and the destination is Hangzhou Xiaoshan Airport. Of course, the user can also leave the starting point blank, in which case the default starting point is the user's current location.

[0060] The starting point can also be the origin and destination specified by the passenger in a ride-hailing scenario. For example, if a passenger takes a ride from Zhejiang University Yuquan Campus to Hangzhou Xiaoshan Airport, the origin is Zhejiang University Yuquan Campus and the destination is Hangzhou Xiaoshan Airport.

[0061] The above application scenarios are merely illustrative examples, and this manual does not limit the application scenarios for route recommendations.

[0062] In this manual, after obtaining the starting point, reachable routes between the starting points can be planned. There are often multiple reachable routes planned, sometimes even dozens or hundreds.

[0063] In one example, all reachable routes between the starting point can be obtained, and then the estimated arrival time for each reachable route can be determined.

[0064] In another example, all reachable routes can be initially filtered, and then subsequent steps can be performed using the reachable routes obtained after the initial filtering, in order to save computational overhead.

[0065] For example, accessible routes can be initially screened based on dimensions such as mileage and cost, thereby filtering out accessible routes that are too long or too expensive.

[0066] For example, all reachable routes can be initially sorted based on dimensions such as mileage and cost, and then the top reachable routes can be obtained.

[0067] In this specification, for each reachable route obtained, the estimated time of arrival (ETA) of the reachable route can be determined. The estimated time of arrival is the estimated time from the origin to the destination, such as 30 minutes, 60 minutes, etc.

[0068] In this specification, an ETA prediction model can be used to predict the estimated arrival time for each reachable route. The ETA prediction model can be a DNN (Deep Neural Networks) model, an XGBoost (eXtremeGradient Boosting) model, or the like.

[0069] The input to the ETA prediction model is typically route information for the reachable route, such as route length, segment identifiers for each road segment along the reachable route, and traffic condition information for each road segment along the reachable route (e.g., smooth traffic, slow traffic, congestion). The output of the ETA prediction model is the estimated arrival time for the reachable route, such as 30 minutes or 60 minutes.

[0070] The ETA prediction model can be trained using sample routes, which are historical routes actually traveled by users. These sample routes have actual arrival time labels, representing the actual time spent by the user on each route. When training the ETA prediction model, the route information of the sample routes can be input into the model to be trained. The model then predicts the estimated arrival time of the sample routes, calculates the difference between the estimated and actual arrival times, and updates the parameters of the ETA prediction model based on this difference for iterative training. To ensure the accuracy of the ETA prediction model, recent sample routes can be used for training, such as those from the most recent month or two months. This specification does not impose any special restrictions on this.

[0071] Step 104: Calculate the goodness rate of the estimated arrival time for each reachable route, where the goodness rate represents the probability that the error of the estimated arrival time is within an acceptable range.

[0072] In this specification, for each reachable route obtained in step 102 above, the goodness rate of the estimated arrival time of the reachable route can also be calculated.

[0073] In this context, "excellent estimated arrival time" means that the difference between the estimated arrival time and the actual arrival time is small and within an acceptable range. For example, the estimated arrival time is greater than or equal to 85% or 90% of the actual arrival time. This range will be referred to as the "excellent interval" in this specification. The "excellent rate" of the estimated arrival time is the probability that the estimated arrival time is excellent, that is, the probability that the estimated arrival time falls within the excellent interval of the actual arrival time. It represents the probability that the error in the estimated arrival time is within an acceptable range and can reflect the accuracy of the estimated arrival time to a certain extent. Generally speaking, a higher "excellent rate" means a higher accuracy of the estimated arrival time, and a lower "excellent rate" means a lower accuracy of the estimated arrival time.

[0074] For example, suppose that when recommending a route, the estimated arrival time is 50 minutes, but it actually takes 55 minutes to reach the destination. Taking the excellent range as [85%, +∞), the estimated arrival time is greater than 85% of the actual arrival time. The error in the estimated arrival time is within an acceptable range and is relatively accurate.

[0075] In one example, the good rate of the expected arrival time can be calculated by assuming that the actual arrival time of the reachable route follows a normal distribution with a mean of its expected arrival time. The specific calculation method will be described in detail in subsequent embodiments.

[0076] In another example, the excellent rate of the sample route can also be used to determine the excellent rate of the predicted arrival time. Still taking the example of starting from the Yuquan Campus of Zhejiang University and arriving at Hangzhou Xiaoshan Airport, the actual arrival time and predicted arrival time of each reachable route in history can be obtained. Then, for each reachable route, calculate the average historical excellent rate of its predicted arrival time, and determine the excellent rate of the predicted arrival time of the corresponding reachable route in this step, etc. This specification places no special restrictions on this.

[0077] Step 106, for each reachable route, determine the time-consuming value of the reachable route according to the predicted arrival time of the reachable route and the excellent rate of the predicted arrival time.

[0078] In this specification, for each reachable route, the weighted summation method can be used to determine the time-consuming value of the reachable route according to the predicted arrival time of the reachable route and the excellent rate of the predicted arrival time.

[0079] In this specification, the non-excellent rate of the predicted arrival time can be calculated according to the excellent rate. The non-excellent rate is the difference between 1 and the excellent rate. The lower the non-excellent rate, the higher the accuracy of the predicted arrival time. In addition, the weight of the predicted arrival time and / or the weight of the non-excellent rate can be preset. When determining the time-consuming value of the reachable route, the preset weight can be obtained first.

[0080] In one example, assume that the weight A of the predicted arrival time (subsequently referred to as the first weight) is preset, 0 < A < 1. After obtaining the weight A of the predicted arrival time, the weight of the non-excellent rate of the predicted arrival time (subsequently referred to as the second weight) can be calculated by 1 - A. Then, the predicted arrival time and the non-excellent rate of the predicted arrival time are weighted and summed, and the value obtained by the weighted summation can be used as the time-consuming value of the reachable route.

[0081] In another example, assume that the second weight of the non-excellent rate of the predicted arrival time is preset. After obtaining the second weight, the first weight of the predicted arrival time can be calculated, and then the predicted arrival time and the non-excellent rate of the predicted arrival time are weighted and summed to obtain the time-consuming value of the reachable route.

[0082] In yet another example, assume that the first weight of the predicted arrival time and the second weight of the non-excellent rate of the predicted arrival time are respectively preset. These two weights can be obtained, and then the weighted summation calculation is performed to further obtain the time-consuming value of the reachable route.

[0083] Of course, in other embodiments, in addition to weighted summation, algorithms such as summation and averaging can also be used to calculate the expected arrival time and the goodness rate of the expected arrival time, thereby obtaining the time score of the corresponding reachable route, etc. This specification does not impose any special restrictions on this.

[0084] Step 108: Determine the reachable routes whose time consumption scores meet the time consumption requirements as the recommended routes between the starting points.

[0085] Based on the aforementioned step 106, after determining the time score for each reachable route, reachable routes whose time scores meet the time requirements can be selected as recommended routes between the starting points.

[0086] For example, the time requirement could be the minimum time score. After determining the time score for each reachable route, the reachable routes can be sorted in ascending order of time score, and then the top N reachable routes can be recommended to the user.

[0087] Taking a self-driving scenario as an example, the top 3 accessible routes can be recommended to the user, that is, the 3 accessible routes with the lowest time scores can be recommended to the user, so that the user can choose a route for navigation.

[0088] Taking a ride-hailing scenario as an example, the top-ranked accessible route can be recommended to the driver, that is, the accessible route with the shortest travel time can be recommended to the driver for navigation.

[0089] As described above, this specification can determine the estimated arrival time and the goodness rate of each reachable route between starting points. Then, based on the estimated arrival time and the goodness rate, it determines the time consumption score of the corresponding reachable route and identifies the recommended route between the starting points based on the time consumption score. By introducing the goodness rate of the estimated arrival time, the accuracy of the time assessment for reachable routes can be improved, minimizing the difference between the actual arrival time and the estimated arrival time of the recommended route, thereby enhancing the accuracy of the recommended route and improving the user experience.

[0090] The following example illustrates the calculation process for the expected arrival time good rate.

[0091] Figure 2 This is a flowchart illustrating an exemplary embodiment of the method for calculating the expected arrival time good rate.

[0092] Please refer to Figure 2 The calculation process for the expected arrival time excellence rate may include the following steps:

[0093] Step 202: Assuming that the actual arrival time of the reachable route follows a normal distribution with a mean equal to its expected arrival time, calculate the standard deviation of the normal distribution.

[0094] In this specification, it is assumed that the actual arrival time of the reachable route follows a normal distribution with a mean equal to the expected arrival time, and the standard deviation of the state distribution can be calculated using a time deviation prediction model.

[0095] The input to the time deviation prediction model is similar to that of the ETA prediction model, and can also be route information of the reachable route, such as route length, segment identifiers of each segment on the reachable route, and traffic condition information of each segment on the reachable route. The output of the time deviation prediction model is the predicted time deviation between the actual arrival time and the estimated arrival time of the reachable route. That is, the time deviation prediction model is used to predict the time deviation between the actual arrival time and the estimated arrival time, for example, 5 minutes, 10 minutes, etc.

[0096] For ease of description, U can be used to represent the time deviation between the actual arrival time and the expected arrival time predicted by the time deviation prediction model, i.e., U = |T 实际 -T ETA |, where T 实际 T represents the actual arrival time. ETA This represents the estimated arrival time.

[0097] In this specification, based on the characteristics of the normal distribution, the standard deviation σ of the normal distribution can be calculated according to the time deviation U, where σ = U / 0.798, that is, the time deviation U is converted into the standard deviation σ.

[0098] The time deviation prediction model is similar to the ETA prediction model and can also be an XGBoost model, a DNN model, etc. The time deviation prediction model can also be trained using historically traveled routes as sample routes, which also have actual arrival time labels. When training the time deviation prediction model, the route information of the sample routes can be input into the model to be trained. The time deviation prediction model then estimates the time deviation U between the actual arrival time and the expected arrival time of the sample routes. 预估 Then, the true time deviation U of the sample route can be calculated based on the actual arrival time label and the expected arrival time of the sample route. 真实 Then it can be based on U 预估 and U 真实 The differences between the time deviation prediction model and the actual deviation prediction model are used to update the parameters of the time deviation prediction model for iterative training.

[0099] Similarly, to ensure the accuracy of the time deviation prediction model, it can be trained using recent sample routes, such as sample routes from the most recent month or the most recent two months. It is worth noting that the sample routes used to train the time deviation prediction model and the sample routes used to train the ETA estimation model can be the same, different, or partially the same and partially different; this specification does not impose any special restrictions on this.

[0100] Step 204: Standardize the normal distribution of the actual arrival time according to the expected arrival time and the standard deviation, so that the actual arrival time follows a standard normal distribution.

[0101] Based on the aforementioned step 202, after calculating the standard deviation σ of the normal distribution, the standard deviation σ and the expected arrival time T can be used as a basis for further calculation. ETA The normal distribution is standardized to transform it into a standard normal distribution, thus obtaining the actual arrival time T. 实际 It follows the transformed standard normal distribution.

[0102] The above standardization conversion process can be expressed by the formula: Y = (T 实际 -T ETA ) / σ.

[0103] Among them, T 实际 T is the actual arrival time of the reachable route. ETA It is the estimated arrival time of the reachable route predicted by the ETA prediction model, σ is the standard deviation calculated in step 202 above, and T is the estimated arrival time. 实际 Follows the mean T ETA The standard deviation of Y is σ, which follows a normal distribution. Y obtained through the above standardization transformation formula follows a standard normal distribution N(0,1).

[0104] Step 206: Under the standard normal distribution, calculate the good rate of the expected arrival time.

[0105] In this specification, the goodness rate of the expected arrival time can be calculated based on the standard normal distribution obtained in step 204 above.

[0106] In this specification, the optimal range for the estimated arrival time can be obtained first during the calculation. It is assumed that the optimal range is [85%, +∞), i.e., the estimated arrival time T... ETA ≥T 实际 ×0.85, and then determine the actual arrival time T based on the excellent range. 实际 The expected interval is obtained as T. 实际 ≤T ETA / 0.85. In other words, the calculation of the expected arrival time good rate is converted into the probability that the actual arrival time falls within the above expected range.

[0107] Since the actual arrival time follows a transformed standard normal distribution, the probability value of the actual arrival time falling within the expected interval can be found using the probability table of the standard normal distribution; that is, T can be obtained by looking up the table. 实际 ≤T ETA The probability value of / 0.85 is used as the good rate of the expected arrival time.

[0108] As can be seen from the above description, this manual assumes that the actual arrival time of reachable routes follows a normal distribution with the mean being their expected arrival time. Based on the characteristics of the normal distribution, the good rate of expected arrival time is calculated, and the implementation process is simple and convenient.

[0109] This manual also provides a navigation method for a ride-hailing scenario.

[0110] Figure 3 This is a flowchart illustrating a navigation method in a ride-hailing scenario, as shown in an exemplary embodiment of this specification.

[0111] Please refer to Figure 3 The navigation method in the ride-hailing scenario may include the following steps:

[0112] Step 302: Obtain the specified starting point.

[0113] In a ride-hailing scenario, passengers can input their starting and ending points through the passenger app. For example, the starting point is Zhejiang University Yuquan Campus, and the destination is Hangzhou Xiaoshan Airport.

[0114] In ride-hailing scenarios, drivers can also input the starting and ending points specified by passengers in the driver's client application; this manual does not impose any special restrictions on this.

[0115] Step 304: Determine the target recommended route between the starting points.

[0116] Based on step 302 above, after obtaining the starting point, the aforementioned method can be used. Figure 1 The route recommendation method shown is to recommend a target route between the starting point and the driver.

[0117] For example, several reachable routes between the starting point can be obtained first, the estimated arrival time of each reachable route and the goodness rate of the estimated arrival time can be determined, and then the time consumption score of the reachable route can be determined based on the estimated arrival time and the goodness rate of the estimated arrival time, and the reachable route with the lowest time consumption score can be determined as the target recommended route.

[0118] Step 306: Recommend the target route to the user so that the user can navigate based on the target route.

[0119] Based on step 304 above, after determining the target recommended route, the target recommended route can be recommended to the driver, for example, by pushing it to the driver's client, which can then use the target recommended route for navigation. Alternatively, the target recommended route can also be recommended to the passenger's client for passengers to view, etc.

[0120] Using the route recommendation method provided in this manual to recommend navigation routes to users can improve the accuracy of time estimation for recommended routes, minimize the difference between the actual arrival time and the estimated arrival time of the recommended route, thereby improving the accuracy of the recommended route and enhancing the user's travel experience.

[0121] Corresponding to the embodiments of the route recommendation method described above, this specification also provides embodiments of a route recommendation device.

[0122] The embodiments of the recommended devices in this specification can be applied in electronic devices. These devices can be implemented through software, hardware, or a combination of both. Taking software implementation as an example, as a logical device, it is formed by the processor of the electronic device loading corresponding computer program instructions from non-volatile memory into memory for execution. From a hardware perspective, such as... Figure 4 The diagram shown is a hardware structure diagram of the electronic device containing the route recommendation device in this specification. (Except for...) Figure 4 In addition to the processor, memory, network interface, and non-volatile memory shown, the electronic device in which the device is located in the embodiment may also include other hardware depending on the actual function of the electronic device, which will not be described in detail here.

[0123] Figure 5 This is a block diagram illustrating an exemplary embodiment of a route recommendation device in this specification.

[0124] Please refer to Figure 5 The route recommendation device can be applied to the aforementioned Figure 4 The electronic device shown can be a server or server cluster deployed with electronic map services, and the route recommendation device includes:

[0125] The route acquisition unit acquires several reachable routes between starting points and determines the estimated arrival time of each reachable route;

[0126] The goodness rate calculation unit calculates the goodness rate of the estimated arrival time for each reachable route, where the goodness rate represents the probability that the error of the estimated arrival time is within an acceptable range;

[0127] The time consumption determination unit determines the time consumption score of each reachable route based on the estimated arrival time and the goodness rate of the estimated arrival time.

[0128] The route recommendation unit determines the reachable routes that meet the time consumption requirements as recommended routes between the starting points.

[0129] Optionally, the step of calculating the goodness rate of the estimated arrival time for each reachable route includes:

[0130] Assuming that the actual arrival time of the reachable route follows a normal distribution with a mean of its expected arrival time, calculate the standard deviation of the normal distribution;

[0131] The normal distribution of the actual arrival time is standardized based on the estimated arrival time and the standard deviation, so that the actual arrival time follows a standard normal distribution.

[0132] Under the standard normal distribution, the good rate of the expected arrival time is calculated.

[0133] Optionally, the step of calculating the standard deviation of the normal distribution, assuming that the actual arrival time of the reachable route follows a normal distribution with a mean equal to its expected arrival time, includes:

[0134] The route information of the reachable route is input into a trained time deviation prediction model, and the time deviation prediction model is used to predict the time deviation between the actual arrival time and the expected arrival time of the reachable route.

[0135] Assuming that the actual arrival time of the reachable route follows a normal distribution with a mean of its expected arrival time, the standard deviation of the normal distribution is calculated based on the time deviation.

[0136] Optionally, the training process of the time deviation prediction model includes:

[0137] Obtain sample routes, which have actual arrival time labels;

[0138] The route information of the sample route is input into the time deviation prediction model to be trained, and the time deviation prediction model is used to predict the time deviation between the actual arrival time and the expected arrival time of the sample route.

[0139] Calculate the true time deviation of the sample route based on the actual arrival time label and the estimated arrival time;

[0140] The time deviation prediction model is trained based on the difference between the time deviation predicted by the time deviation prediction model and the actual time deviation.

[0141] Optionally, the step of calculating the good / good rate of the expected arrival time under the standard normal distribution includes:

[0142] The expected range of the actual arrival time is determined based on the favorable range of the estimated arrival time;

[0143] The probability value that the actual arrival time falls within the expected interval under the standard normal distribution is calculated as the good rate of the expected arrival time.

[0144] Optionally, the time requirement is to have the lowest time score, and the step of determining the time score of the reachable route based on the estimated arrival time and the goodness rate of the estimated arrival time includes:

[0145] The non-excellent rate of the estimated arrival time is calculated based on the excellent rate of the estimated arrival time;

[0146] Determine a first weight for the estimated arrival time and a second weight for the non-excellent rate;

[0147] Based on the first weight and the second weight, the estimated arrival time and the non-excellent rate are weighted and summed to obtain the time score of the reachable route.

[0148] The specific implementation process of the functions and roles of each unit in the above device can be found in the implementation process of the corresponding steps in the above method, and will not be repeated here.

[0149] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of the solution in this specification according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0150] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer, which can take the form of a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email sending and receiving device, game console, tablet computer, wearable device, or any combination of these devices.

[0151] In a typical configuration, a computer includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0152] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0153] Computer-readable media, including both permanent and non-permanent, removable and non-removable media, can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, disk storage, quantum memory, graphene-based storage media or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0154] Corresponding to the embodiments of the aforementioned route recommendation method, this specification also provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the following steps:

[0155] Obtain several reachable routes between starting points and determine the estimated arrival time for each of the reachable routes;

[0156] Calculate the goodness rate of the estimated arrival time for each reachable route, where the goodness rate represents the probability that the error of the estimated arrival time is within an acceptable range;

[0157] For each reachable route, a time score is determined based on the estimated arrival time and the goodness rate of the estimated arrival time.

[0158] The reachable routes that meet the time consumption requirements are determined as the recommended routes between the starting points.

[0159] Optionally, the step of calculating the goodness rate of the estimated arrival time for each reachable route includes:

[0160] Assuming that the actual arrival time of the reachable route follows a normal distribution with a mean of its expected arrival time, calculate the standard deviation of the normal distribution;

[0161] The normal distribution of the actual arrival time is standardized based on the estimated arrival time and the standard deviation, so that the actual arrival time follows a standard normal distribution.

[0162] Under the standard normal distribution, the good rate of the expected arrival time is calculated.

[0163] Optionally, the step of calculating the standard deviation of the normal distribution, assuming that the actual arrival time of the reachable route follows a normal distribution with a mean equal to its expected arrival time, includes:

[0164] The route information of the reachable route is input into a trained time deviation prediction model, and the time deviation prediction model is used to predict the time deviation between the actual arrival time and the expected arrival time of the reachable route.

[0165] Assuming that the actual arrival time of the reachable route follows a normal distribution with a mean of its expected arrival time, the standard deviation of the normal distribution is calculated based on the time deviation.

[0166] Optionally, the training process of the time deviation prediction model includes:

[0167] Obtain sample routes, which have actual arrival time labels;

[0168] The route information of the sample route is input into the time deviation prediction model to be trained, and the time deviation prediction model is used to predict the time deviation between the actual arrival time and the expected arrival time of the sample route.

[0169] Calculate the true time deviation of the sample route based on the actual arrival time label and the estimated arrival time;

[0170] The time deviation prediction model is trained based on the difference between the time deviation predicted by the time deviation prediction model and the actual time deviation.

[0171] Optionally, the step of calculating the good / good rate of the expected arrival time under the standard normal distribution includes:

[0172] The expected range of the actual arrival time is determined based on the favorable range of the estimated arrival time;

[0173] The probability value that the actual arrival time falls within the expected interval under the standard normal distribution is calculated as the good rate of the expected arrival time.

[0174] Optionally, the time requirement is to have the lowest time score, and the step of determining the time score of the reachable route based on the estimated arrival time and the goodness rate of the estimated arrival time includes:

[0175] The non-excellent rate of the estimated arrival time is calculated based on the excellent rate of the estimated arrival time;

[0176] Determine a first weight for the estimated arrival time and a second weight for the non-excellent rate;

[0177] Based on the first weight and the second weight, the estimated arrival time and the non-excellent rate are weighted and summed to obtain the time score of the reachable route.

[0178] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0179] The above description is merely a preferred embodiment of this specification and is not intended to limit this specification. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this specification should be included within the scope of protection of this specification.

Claims

1. A route recommendation method, comprising: Obtain several reachable routes between the starting points and determine the estimated arrival time for each of the reachable routes; Calculate the goodness rate of the estimated arrival time for each reachable route, where the goodness rate represents the probability that the error of the estimated arrival time is within an acceptable range; Calculate the non-excellent rate of the estimated arrival time based on the excellent rate of the estimated arrival time; determine a first weight for the estimated arrival time and a second weight for the non-excellent rate; and perform a weighted summation of the estimated arrival time and the non-excellent rate based on the first weight and the second weight to obtain the time score of the reachable route. The reachable routes that meet the time consumption requirements are determined as the recommended routes between the starting points.

2. The method according to claim 1, wherein calculating the goodness rate of the estimated arrival time for each reachable route includes: Assuming that the actual arrival time of the reachable route follows a normal distribution with a mean of its expected arrival time, calculate the standard deviation of the normal distribution; The actual arrival time is standardized by applying the expected arrival time and the standard deviation to the normal distribution, so that the actual arrival time follows a standard normal distribution. Under the standard normal distribution, the good rate of the expected arrival time is calculated.

3. The method according to claim 2, wherein calculating the standard deviation of the normal distribution, assuming that the actual arrival time of the reachable route follows a normal distribution with a mean equal to its expected arrival time, includes: The route information of the reachable route is input into a trained time deviation prediction model, and the time deviation prediction model is used to predict the time deviation between the actual arrival time and the expected arrival time of the reachable route. Assuming that the actual arrival time of the reachable route follows a normal distribution with a mean of its expected arrival time, the standard deviation of the normal distribution is calculated based on the time deviation.

4. The method according to claim 3, wherein the training process of the time deviation prediction model includes: Obtain sample routes, which have actual arrival time labels; The route information of the sample route is input into the time deviation prediction model to be trained, and the time deviation prediction model is used to predict the time deviation between the actual arrival time and the expected arrival time of the sample route. Calculate the true time deviation of the sample route based on the actual arrival time label and the estimated arrival time; The time deviation prediction model is trained based on the difference between the time deviation predicted by the time deviation prediction model and the actual time deviation.

5. The method according to claim 2, wherein calculating the good rate of the expected arrival time under the standard normal distribution comprises: The expected range of the actual arrival time is determined based on the favorable range of the estimated arrival time; The probability value that the actual arrival time falls within the expected interval under the standard normal distribution is calculated as the good rate of the expected arrival time.

6. A navigation method for a ride-hailing scenario, comprising: Get the specified starting point; The route recommendation method according to any one of claims 1-5 determines the target recommended route between the starting points; The target recommended route is recommended to the user so that the user can navigate based on the target recommended route.

7. A route recommendation device, comprising: The route acquisition unit acquires several reachable routes between starting points and determines the estimated arrival time of each reachable route; The goodness rate calculation unit calculates the goodness rate of the estimated arrival time for each reachable route, where the goodness rate represents the probability that the error of the estimated arrival time is within an acceptable range; The time consumption determination unit calculates the non-excellent rate of the estimated arrival time based on the excellent rate of the estimated arrival time; determines a first weight of the estimated arrival time and a second weight of the non-excellent rate; and performs a weighted summation of the estimated arrival time and the non-excellent rate based on the first weight and the second weight to obtain the time consumption score of the reachable route. The route recommendation unit determines the reachable routes that meet the time consumption requirements as recommended routes between the starting points.

8. An electronic device, comprising: processor; Memory used to store processor-executable instructions; The processor implements the method as described in any one of claims 1-6 by executing the executable instructions.

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

Images