Methods, related apparatus and servers for determining the estimated arrival time of a route

Abstract

This disclosure provides a method, related apparatus, and server for determining the estimated arrival time of a route. The method, for road segment units (including links or links and their entry points) included in the route, performs the following steps: based on the obtained estimated arrival time of the current link in the current road segment unit, the obtained uncertainty parameter of the estimated arrival time of the current link, and the feature set of the current road segment unit, at least the estimated travel time of the current road segment unit is obtained; the sum of the estimated arrival time of the current link and the estimated travel time of the current road segment unit is obtained; if the current link is the link of the last road segment unit of the route, the sum is determined as the estimated arrival time of the route. This disclosure makes the final estimated travel time of each road segment unit more accurate, and the corresponding estimated arrival time of the entire navigation planning route more accurate.

Description

Technical Field

[0001] This disclosure relates to the field of navigation technology, and in particular to a method, related apparatus and server for determining the estimated arrival time of a route. Background Technology

[0002] In the ride-hailing service sector, when a user plans a route given a starting point A and a destination B, the ride-hailing app typically needs to display the estimated time of arrival (ETA) for the relevant route. For example, starting from point A to point B, if you depart now and take a navigation route that avoids traffic congestion, you are expected to arrive at point B in 15 minutes.

[0003] Existing techniques use baseline models to predict the ETA of a route. Baseline models are typically modeled using the travel time and intersection compensation time of at least one link within the route, and are expressed as follows:

[0004]

[0005] Here, link time refers to the travel time of a single link in the route. The travel time of a single link can be obtained by fusing the historical average travel time and real-time travel time of that link. The fusing method is shown in the following formula:

[0006] Link time = k * SP t +(1-k)*Autolr t

[0007] Among them, SP t This is the historical average travel time of the link, calculated based on historical travel data, and reflects the regularity of the link's travel time; Autolr t It is the real-time access time of the link.

[0008] As can be seen from the above model, although the existing technology takes into account the historical average travel time of links in the route, the real-time travel time, and the intersection compensation time in the route, the above basic model does not fully express the factors affecting ETA due to the complexity of road traffic conditions. Therefore, the existing technology has the problem of inaccurate ETA prediction. Summary of the Invention

[0009] In view of the above problems, this disclosure is made in order to provide a method, related apparatus and equipment for determining the estimated arrival time of a route to overcome or at least partially solve the above problems.

[0010] In a first aspect, embodiments of this disclosure provide a method for determining the estimated arrival time of a route, the route comprising at least one road segment unit, the road segment unit comprising a road link or a road link and its entry point, wherein the following steps are performed on the road segment unit:

[0011] Based on the estimated arrival time of the current link in the current road segment unit, the uncertainty parameter of the estimated arrival time of the current link, and the feature set of the current road segment unit, at least the estimated travel time of the current road segment unit can be obtained.

[0012] Obtain the sum of the estimated arrival time of the current link and the estimated travel time of the current road segment unit;

[0013] If the current link is the link of the last segment unit of the route, then the sum is determined as the estimated arrival time of the route.

[0014] In one or more embodiments, if the current link is not the link of the last segment unit of the route, the method further includes:

[0015] The sum is determined as the estimated arrival time of the link in the next segment unit, where the next segment unit is the segment unit that the route enters after passing the current segment unit.

[0016] Based on the obtained estimated arrival time of the current link, the obtained uncertainty parameter of the estimated arrival time of the current link, and the feature set of the current segment unit, the uncertainty parameter of the estimated arrival time of the link in the next segment unit is obtained.

[0017] Take the link in the next road segment unit as the current link and return to the step of obtaining at least the estimated travel time of the current link.

[0018] In one or more embodiments, based on the obtained estimated arrival time of the current link, the obtained uncertainty parameter of the estimated arrival time of the current link, and the feature set of the current segment unit, the uncertainty parameter of the estimated arrival time of the link in the next segment unit is obtained, including:

[0019] The estimated arrival time of the current link, the uncertainty parameters of the estimated entry time of the current link, and the feature set of the current road segment unit are input into a pre-trained recurrent neural network model. The recurrent neural network model outputs the estimated passage time of the current road segment unit and the uncertainty parameters of the estimated entry time of the current link in the current road segment unit.

[0020] In one or more embodiments, the feature set of the road segment unit includes: a discrete feature set and a continuous feature set;

[0021] The discrete feature set includes at least: road type features of links in the road segment unit, road congestion status features of links in the road segment unit, type features of intersections in the road segment unit, and navigation action features of the route at intersections in the road segment unit;

[0022] The continuous feature set includes at least: at least one historical average travel time of a link in the road segment unit, and the real-time travel time of a link in the road segment unit; the at least one historical average travel time includes: each historical average travel time corresponding to multiple historical arrival times before and / or after the estimated arrival time of the current link.

[0023] In one or more embodiments, the method further includes:

[0024] The features in the discrete feature set are input into the embedding layer, and the output of the embedding layer is used as the input of the recurrent neural network model.

[0025] In one or more embodiments, the uncertainty parameter is used to characterize the time deviation uncertainty value of the expected entry time of the link in the current segment unit.

[0026] In a first aspect, embodiments of this disclosure provide an apparatus for determining the estimated arrival time of a route, comprising:

[0027] The estimated travel time acquisition module is used to obtain at least the estimated travel time of the current road segment unit based on the estimated arrival time of the current link in the current road segment unit, the uncertainty parameter of the estimated arrival time of the current link, and the feature set of the current road segment unit; the route includes at least one road segment unit, and the road segment unit includes a road segment link or a road segment link and its entry intersection;

[0028] The time accumulation module is used to obtain the sum of the estimated arrival time of the current link and the estimated travel time of the current road segment unit;

[0029] The judgment module is used to determine whether the current link is the link of the last segment unit of the route;

[0030] An estimated arrival time determination module is used to determine the estimated arrival time of the route by summing the output of the time accumulation unit when the determination module determines that the link is the last segment unit of the route.

[0031] This disclosure also provides a computer-readable storage medium storing computer instructions that are executed by a processor as described above for determining the estimated arrival time of a route.

[0032] This disclosure also provides a navigation server, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described above for determining the estimated arrival time of a route.

[0033] The beneficial effects of the above-described technical solutions provided in this disclosure include at least the following:

[0034] The method, apparatus, and related server for determining the estimated arrival time of a route provided in this disclosure, for each road segment unit in the route, based on the obtained estimated arrival time of the current link in the current road segment unit, the obtained uncertainty parameter of the estimated arrival time (entry time) of the current link, and various feature sets of the road segment unit, can obtain the estimated travel time of the current road segment unit, and then determine the sum of the estimated arrival time of the current link and the estimated travel time of the current road segment unit; if the current link is the last link of the route, then the sum is the estimated arrival time of the route. This disclosure fully utilizes... The calculation of the estimated travel time of each link in each road segment unit is influenced by the estimated arrival time of the previous link and the uncertainty of the estimated arrival time of the previous link. Compared with the existing technology, this method fully considers the actual situation in which the estimated arrival time of the previous link and the uncertainty of the estimated arrival time of the previous link in the actual travel process will affect the estimated travel time of the links in the adjacent downstream road segment units. This makes the estimated travel time of each link in the final navigation planning video more accurate, and the estimated arrival time of the entire planned navigation route is also more accurate.

[0035] This embodiment employs a pre-trained neural network model to calculate the estimated travel time for each road segment unit in the route and to learn the estimated arrival time uncertainty parameters of links. Since this neural network model can be trained on real user travel data samples to obtain the weight relationship between the historical average travel times corresponding to multiple times before and after a certain time when entering a certain link at different times, the accuracy of ETA calculation can be further improved by learning the estimated travel time of road segment units through this neural network.

[0036] Other features and advantages of this disclosure will be set forth in the following description and will be apparent in part from the description or may be learned by practicing the disclosure. The objects and other advantages of this disclosure may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings.

[0037] The technical solutions of this disclosure will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0038] The accompanying drawings are provided to further illustrate the present disclosure and form part of the specification. They are used together with the embodiments of the present disclosure to explain the disclosure and do not constitute a limitation thereof. In the drawings:

[0039] Figure 1 A flowchart illustrating a method for determining the estimated arrival time of a route, provided as an embodiment of this disclosure;

[0040] Figure 2 A schematic diagram of a navigation route planning provided in an embodiment of this disclosure;

[0041] Figure 3 A schematic diagram illustrating the model composition for determining the estimated travel time using a neural network model, as provided in this embodiment of the disclosure;

[0042] Figure 4 This is a schematic diagram of a device for determining the estimated arrival time of a route, provided as an embodiment of the present disclosure. Detailed Implementation

[0043] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0044] The inventors of this disclosure have discovered that existing baseline models for calculating ETA are overly simplistic. Due to the uncertainty of the estimated arrival time (i.e., estimated entry time) of each link in the route, the selection of the historical average travel time and real-time travel time for each link may be uncertain. Furthermore, the impact of the uncertainty of the travel time of upstream links in the navigation route on the travel time of downstream links (historical average travel speeds may differ at different times) is not considered, leading to inaccurate ETA calculations for the final route. Therefore, to address the problem of inaccurate ETA calculations, embodiments of this disclosure start from each road segment unit in the navigation route, fully considering the uncertainty of the estimated arrival time (estimated entry time) of each link in the road segment unit and the impact of the uncertainty of the travel time of links in upstream road segment units on the uncertainty of the estimated travel time in downstream road segment units (including the estimated travel time of links and the estimated time of entry into intersections). This provides a method, related apparatus, and server for determining the estimated arrival time of a route, thereby improving the accuracy of ETA estimation.

[0045] This disclosure describes two types of road segment units: one type includes only links, and the other type includes links and entrances to those links, where an entrance to a link is the intersection accessed through that link. Typically, roads in electronic maps are divided into several links, where a link is shorter than the road. A link located in the middle of a road enters another link. For roads connecting intersections, the last link of that road will have an entrance. This disclosure treats such links as a single unit.

[0046] The following provides a detailed description of the specific implementation methods, related apparatus, and server for determining the estimated arrival time of a route, as provided in the embodiments of this disclosure.

[0047] This disclosure provides a method for determining the estimated arrival time of a route, which processes each link or link and its entry point in the route as a whole, i.e., a road segment unit. For each road segment unit, the execution flow is as follows: Figure 1 As shown, the following steps may be included:

[0048] S101. Based on the estimated arrival time of the current link in the current road segment unit, the uncertainty parameter of the estimated arrival time of the current link, and the feature set of the current road segment unit, at least the estimated travel time of the current road segment unit is obtained.

[0049] In the embodiments of the invention, the route can be a navigation planning route that has been generated in various ways.

[0050] In step S101 above, the route is composed of several road segment units. Each road segment unit contains a link and a subsequent entry point. For each road segment unit in the route, the estimated arrival time (also the estimated entry time) of the current link in the current road segment unit is obtained. This refers to the estimated time to reach the current link from the link where the starting point of the route is located. The "entry point" mentioned above refers to the intersection that immediately follows the current link in the route.

[0051] The uncertainty parameter of the estimated arrival time of a link in each road segment unit is used to characterize the time deviation (early deviation or late deviation, that is, whether it is earlier or later than the estimated arrival time) of the link within a preset range.

[0052] For the link in the first segment unit of the route, its estimated arrival time is determined (the estimated arrival time is the estimated entry time, which can be determined based on the departure time), so the uncertainty parameter of the estimated arrival time of the corresponding current link is zero.

[0053] For each subsequent link in a non-first segment unit, the uncertainty parameter of its estimated arrival time may be affected by the uncertainty parameter of the estimated arrival time of the link in the upstream segment unit. Therefore, it can be obtained through neural network learning.

[0054] S102. Obtain the sum of the estimated arrival time of the current link and the estimated travel time of the current road segment unit;

[0055] S103. Determine whether the current link is the link of the last segment unit of the route. If so, execute the following step S104.

[0056] S104. The sum of the estimated arrival time of the current link and the estimated travel time of the current road segment unit is determined as the estimated arrival time of the route.

[0057] In step S104, if the current link is the link in the last road segment unit, the estimated travel time of the current road segment unit is added to the estimated arrival time of the last link to obtain the estimated arrival time of the route.

[0058] The method, apparatus, and related server provided in this disclosure for determining the estimated arrival time of a route, for each road segment unit (containing a link, or containing a link and its entry point) in the route, based on the obtained estimated arrival time of the current link in the current road segment unit, the obtained uncertainty parameter for the estimated arrival time (entry time) of the current link, and various feature sets of the current road segment unit, can obtain the estimated travel time of the current road segment unit, and then determine the sum of the estimated arrival time of the current link and the estimated travel time of the current road segment unit; if the current link is the link of the last road segment unit of the route, then the sum is the estimated arrival time of the route. The example fully utilizes the estimated arrival time of the previous link (i.e., the link in the previous segment unit) and the uncertainty of the estimated arrival time of the previous link in each segment unit to influence the calculation of the estimated travel time of the current segment unit. Compared with the prior art, because it fully considers the actual situation that the estimated arrival time of the link in the previous segment unit and the uncertainty of the estimated arrival time of the link in the previous segment unit will affect the estimated travel time of the adjacent downstream link during actual travel, the estimated travel time of the link in each segment unit and its entry point in the final navigation planning route is more accurate, and the estimated arrival time of the entire planned navigation route is also more accurate.

[0059] In the method for determining the estimated arrival time of a route provided in the embodiments of this disclosure, the estimated travel time of the current road segment unit is learned as a whole to obtain the estimated arrival time of the current link, the uncertainty parameter of the estimated arrival time of the current link, and the feature set of the current road segment unit. The estimated travel time of the current road segment unit and the uncertainty parameter for the estimated arrival time of the next link are obtained, and then the calculation is performed iteratively to deduce the estimated arrival time of the next link. Based on the estimated arrival time of the next link and the uncertainty parameter of the estimated arrival time of the next link (the calculation result of the previous link), the estimated travel time of the next road segment unit and the uncertainty parameter for the estimated arrival time of the link in the next road segment unit are obtained, until the link in the last road segment unit. Since there are no other links after the last link, it is not necessary to calculate the uncertainty parameter of the estimated arrival time of the link in the next road segment unit for the last link.

[0060] In one embodiment, the uncertainty parameter is used to characterize the uncertainty value of the time deviation of the obtained expected arrival time of the current link (i.e., the expected arrival time).

[0061] Based on this, for each link in the route that is not the last segment unit, in addition to performing the above steps S101 to S103, the following steps are still referred to: Figure 1 As shown, the following steps can also be performed:

[0062] S105. The sum of the estimated arrival time of the current link and the estimated travel time of the current road segment unit is used as the estimated arrival time to the link in the next road segment unit;

[0063] The next segment unit is the segment unit that the route enters after passing through the current segment unit;

[0064] S106. Based on the obtained estimated arrival time of the current link, the obtained uncertainty parameter of the estimated arrival time of the current link, and the feature set of the current segment unit, obtain the uncertainty parameter of the estimated arrival time of the link in the next segment unit.

[0065] S107. Take the link in the next road segment unit as the current link, and then proceed to step S101 above.

[0066] In steps S105 to S107, based on the estimated arrival time of the current link in the current road segment unit, the uncertain parameters of the estimated arrival time of the current link that have been obtained, and the feature set of the current road segment unit, the uncertain parameters of the estimated arrival time of the link in the next road segment unit (i.e., the next link) can be obtained.

[0067] Specifically, in steps S101 and S106 above, the uncertainty parameters of the estimated passage time of the current link and the estimated arrival time of the next link of the current link can be obtained through neural network model learning.

[0068] For example, the following method can be used: input the obtained estimated arrival time of the current link, the obtained uncertainty parameter of the estimated arrival time of the current link, and the feature set of the current road segment unit into a pre-trained recurrent neural network model, which outputs the estimated travel time of the current road segment unit and the uncertainty parameter of the estimated arrival time of the current link.

[0069] During training, the pre-trained neural network model uses a large amount of user traffic data as training samples. This user traffic data includes the actual travel time of users from a set starting point to a set ending point. These sample data implicitly contain the relationship between the entry time of each link and the travel time of the road segment unit to which that link belongs. Using such training samples, the neural network model can learn the relationship between the entry time and the travel time of the road segment unit to which the link belongs.

[0070] The aforementioned neural network model can employ various existing models, such as CNN convolutional neural networks, RNN recurrent neural networks, and DNN deep neural networks, and this disclosure does not impose any specific limitations on them.

[0071] In one embodiment, the feature set of the current road segment unit in steps S101 and S106 above may, for example, include: a discrete feature set and a continuous feature set; wherein:

[0072] 1. The discrete feature set shall include at least: road type features of links in road segment units, road congestion status features of links in road segment units, intersection type features of intersections in road segment units, and navigation action features of routes at intersections in road segment units;

[0073] The road type characteristics of the current link include, but are not limited to: such as highways, expressways, main roads, ring roads, national highways, provincial highways, etc.

[0074] The characteristics of intersection types include, but are not limited to: for example, intersections without traffic lights, T-junctions with traffic lights, and crossroads with traffic lights;

[0075] Navigation action characteristics refer to the turning type characteristics of the planned navigation route, including but not limited to: such as left turn, right turn, straight ahead, etc.

[0076] The current road congestion status characteristics of a link include, but are not limited to: moderate congestion (e.g., medium congestion level), severe congestion (e.g., high congestion level), and free flow (e.g., low congestion level).

[0077] 2. The continuous feature set shall include at least: at least one historical average travel time of a link in the road segment unit, and the real-time travel time of a link in the road segment unit;

[0078] At least one historical average travel time, specifically including: the historical average travel time corresponding to multiple historical arrival times before and / or after the estimated arrival time of the current link; for example, the historical average travel time corresponding to multiple historical arrival times included in a preset range before and / or after the estimated arrival time of the current link.

[0079] The real-time travel time of the current link is the sum of the user's actual departure time and the estimated travel time of all road segments preceding each link. The actual entry time of each link is determined, and the real-time travel time of the current link can be obtained based on the actual entry time.

[0080] In this embodiment of the disclosure, the historical average travel time corresponding to multiple times within a preset range before and / or after the estimated entry time of the current link can be, for example, the historical average travel time within one hour before and after the estimated entry time of the current link (in this embodiment, each time represents a preset small unit time period, such as 10 minutes). For example, if the estimated entry time of the current link is between 10:00 AM and 10:10 AM, then the historical average travel time corresponding to each unit time period within the two-hour period from 9:00 AM to 11:10 AM is taken.

[0081] Historical average travel time, which can be derived from extensive statistical analysis of historical travel data, represents the average time required to traverse a given link within a specific timeframe. Since traffic conditions vary for the same link on different dates and at different times of day, the travel time will also differ, necessitating the calculation of averages based on these variations. For instance, for a given link, Monday is typically the first day of work, resulting in more congestion than Wednesday; and Monday morning at 7:00 AM is more congested than at 2:00 PM due to the morning rush hour. The historical average travel time for each timeframe can be calculated by averaging historical travel times for the same period over multiple days.

[0082] In one embodiment, the discrete feature set is not directly input into the neural network model. Instead, the various features in the discrete feature set are first input into the embedding layer, and then the output of the embedding layer is used as the input of the recurrent neural network model.

[0083] The method for determining the estimated arrival time of a route provided in the embodiments of this disclosure is illustrated by a practical example.

[0084] Starting from the first link in the first segment unit of the navigation route, for each current link, the estimated arrival time of the current link, the estimated entry time uncertainty parameter of the current link (calculated from the previous segment unit), and the feature set information of the current segment unit can be calculated based on the departure time and the cumulative value of the estimated travel time of all segment units before the current link. This allows us to determine the estimated travel time of the current segment unit and further obtain the estimated entry time uncertainty parameter of the link in the next segment unit of the current segment unit.

[0085] The estimated travel time of the current road segment unit is then added to the estimated travel time of the departure time and all the estimated travel times of the preceding road segment units to obtain the estimated arrival time of the link in the next road segment unit. Then, for the link in the next road segment unit, the above process is repeated until the link in the last road segment unit of the route, that is, the last link. The difference is that the last link does not need to calculate the estimated travel time of the next link.

[0086] The following is in conjunction with the appendix Figure 2 and Figure 3 The examples shown further illustrate the embodiments of this disclosure:

[0087] Reference Figure 2 As shown, the computer system can determine the navigation route based on the user-defined starting point A and ending point B, and then obtain the characteristic information of each link (link1 to link8) included in the navigation route, as well as their entry points and each road segment unit.

[0088] To facilitate calculation, the user's departure time at starting point A is set to 0:00. Then, the estimated arrival time of each subsequent link can be obtained by accumulating the estimated travel time of all preceding road segments.

[0089] Reference Figure 3 As shown, starting from the link of the first segment unit of the navigation planned route, for a link of a certain nth segment unit, the cumulative value of the estimated travel time of all segment units before the link of the nth segment unit is ∑y. n-1 This refers to the estimated arrival time of the current link, the uncertainty parameters of the estimated arrival time of the current link (i.e., the uncertainty parameters Z1 and Z2 calculated using the link in the previous road segment unit, i.e., the (n-1)th link), and the feature vector set {f} corresponding to the feature information set of this road segment unit (including relevant feature information of the link and relevant feature information of the intersection). nThe input is fed into a pre-trained neural network model, which outputs the estimated travel time y of the link in the nth road segment unit. n And determine the uncertainty parameters Z1 and Z2 for the estimated arrival time of the link in the next segment unit (i.e., the (n+1)th segment unit) of the nth segment unit.

[0090] It should be noted that for the link in the first road segment unit, its expected arrival time, i.e. departure time, is definite. Therefore, during the calculation, the expected arrival time input into the neural network model is 0, and the uncertainty parameters Z1 and Z2 of the expected arrival time of the current link are both 0.

[0091] The estimated travel time y of the link of the nth road segment unit will be obtained. n The estimated travel time of the first n-1 road segments ∑y n-1 By adding them together, we can obtain the cumulative value of the estimated travel time ∑y for the first n road segment units. n That is, the estimated arrival time of the link in the (n+1)th segment unit (since the departure time is 0, then ∑y n That is, the estimated arrival time of the (n+1)th link). Then repeat the above process, and ∑y n The uncertainty parameters Z1 and Z2 of the estimated arrival time of the link in the (n+1)th road segment unit, and the set of feature vectors corresponding to the relevant feature information set of the (n+1)th road segment unit {f n+1 The value is input to the neural network model, and the neural network model outputs the estimated travel time y for the (n+1)th road segment unit. n+1 The estimated arrival time of the next link in the next segment unit (i.e., the (n+2)th link) is calculated, along with the uncertainty parameters Z1 and Z2. This process continues until the cumulative estimated travel time of all segments in the navigation route is calculated, ultimately yielding the estimated arrival time of the entire route.

[0092] The aforementioned uncertainty parameters Z1 and Z2 represent the uncertainty (i.e., time deviation) of the expected arrival time of the corresponding link within different time ranges, specifically:

[0093] Z1 is the uncertainty parameter of the link within a preset time range before the expected entry time;

[0094] Z2 is the uncertainty parameter of the link within a preset time range after the expected entry time;

[0095] For example, before entering a link, it's uncertain whether the actual arrival time will be earlier or later than the expected arrival time. Therefore, two parameters, Z1 and Z2, are set to represent the uncertainty of the time deviation within a preset range before and after the expected arrival time, respectively. For instance, if the expected arrival time for the third link (L3) is 10:00 AM, Z1 represents the uncertainty of the time deviation within one hour before the expected arrival time of L3 at 10:00 AM; Z2 represents the uncertainty of the time deviation within one hour after 10:00 AM.

[0096] During training, the pre-trained neural network model uses a large amount of user traffic data as training samples. This user traffic data includes the actual travel time of users from a set starting point to a set ending point. These sample data contain the relationship between the entry time of each link and the actual travel time of the road segment unit to which that link belongs. Using such training samples, the weights of the historical average travel times corresponding to multiple times within a preset range before and / or after any expected arrival time when entering a link are trained.

[0097] The feature information of any road segment unit (including links or links and entry points) used as input to the neural network model may include one or more of the following:

[0098] The historical average travel time corresponding to multiple times within a preset range before and / or after the expected entry time of the link in the current road segment unit;

[0099] The road type to which the link belongs in the current road segment unit, such as highway, provincial highway, national highway, expressway, etc.

[0100] Real-time road traffic information of the link in the current road segment unit, such as real-time road travel time, real-time road speed, and road congestion status (general congestion, extreme congestion, and smooth traffic).

[0101] The characteristics of the entrance intersection type of the link in the current road segment unit, such as T-shaped intersections with traffic lights, cross-shaped intersections, etc.;

[0102] The navigation action characteristics of the link entering the intersection in the current road segment unit, such as left turn, right turn, and straight ahead.

[0103] Specifically, starting from the starting point A of the navigation route, determine the link (n=1) of the first road segment unit, i.e., the estimated travel time of L1. Based on... Figure 3 The model shown has the following inputs for the neural network model that calculates the L1 estimated travel time:

[0104] 1.1) The cumulative estimated travel time ∑y for all road segment units before L1 n-1 This refers to the estimated arrival time of the link (departure time is set to 0); since L1 is the first link in the road segment unit, and there are no links before it, the cumulative value of the estimated travel time of all road segment units before L1 (n=1) ∑y0 is zero.

[0105] 1.2) The uncertainty parameters Z1 and Z2 of the estimated travel time of the previous link of L1 are the same as the uncertainty parameters of the estimated arrival time of L1. Since L1 is the first link of the road segment unit and there are no links before it, the uncertainty parameters of the estimated arrival time of L1 are also zero.

[0106] 1.3) Determine the feature vector set {f1} corresponding to the feature information of the road segment unit where L1 is located, including: the road attribute features to which L1 belongs, the real-time road traffic information features of L1, the historical average travel time features corresponding to multiple time periods (with 10 minutes as a basic time unit) within the preset range before and / or after the expected entry time of L1, the type features of L1 entering the intersection and the navigation action features, etc.

[0107] For example, if the expected entry time for L1 is 10:00 AM, then we select the historical average travel time characteristics for each time period (10 minutes per time period) between 9:00 AM and 10:00 AM, one hour before 10:00 AM. We also select the historical average travel time characteristics for each time period (10 minutes per time period) between 10:10 AM and 11:10 AM, one hour after 10:00 AM and 10:10 AM. Adding the historical average travel time characteristics for the time period between 10:00 AM and 10:10 AM, we have a total of 13 historical average travel time characteristics.

[0108] The above parameters are input into the neural network model. Based on the information contained in the input ∑y0, Z1 and Z2 and {f1}, the neural network model outputs the estimated travel time y1 of the road segment unit where L1 is located and the uncertainty parameters Z11 and Z21 for the estimated entry time of L2.

[0109] By analogy, the above neural network model is used to calculate the uncertainty parameters of the estimated travel time of the road segment unit where L2 is located and the estimated arrival time of L3 for L2. The calculation is repeated until the estimated travel time of the road segment unit where the last Link, i.e. Link8, is located is calculated. Then, the estimated arrival time of Link8 is summed with the estimated arrival time of L8 to obtain the final arrival time of the entire route.

[0110] It should be noted that, in this embodiment of the disclosure, the estimated travel time of a link and the estimated travel time of an intersection can also be learned separately using a neural network model. Finally, the estimated arrival time of the entire route is obtained by adding the cumulative values ​​of the estimated travel times of all links and the cumulative values ​​of the estimated travel times of intersections. The specific implementation method for learning the estimated travel times of links and intersections is similar to the aforementioned implementation method for learning the estimated travel times of road segment units using a neural network. The difference is that, during the learning process, the input features are the relevant features of links and the relevant features of intersections, which will not be described in detail here.

[0111] Based on the same inventive concept, this disclosure also provides an apparatus and related server for determining the estimated arrival time of a route. Since the principle by which these apparatuses and servers solve the problem is similar to the aforementioned method for determining the estimated arrival time of a route, and the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, it will not be elaborated here.

[0112] This disclosure also provides an apparatus for determining the estimated arrival time of a route, referring to... Figure 4 As shown, the device includes:

[0113] The estimated travel time acquisition module 51 is used to obtain at least the estimated travel time of the current road segment unit based on the estimated arrival time of the current link in the current road segment unit, the uncertainty parameter of the estimated arrival time of the current link, and the feature set of the current road segment unit; the route includes at least one road segment unit, and the road segment unit includes a road segment link or a road segment link and its entry intersection;

[0114] The time accumulation module 52 is used to obtain the sum of the estimated arrival time of the current link and the estimated travel time of the current road segment unit;

[0115] The judgment module 53 is used to determine whether the current link is the link of the last segment unit of the route;

[0116] The estimated arrival time determination module 54 is used to determine the estimated arrival time of the route by summing the output value of the time accumulation unit when the determination module determines that the link is the last segment unit of the route.

[0117] In one embodiment, the apparatus for determining the estimated arrival time of a route further includes: an uncertainty parameter acquisition module 55; wherein:

[0118] The estimated arrival time determination module 54 is further configured to determine the sum as the estimated arrival time of the link in the next road segment unit, wherein the next road segment unit is the road segment unit that the route enters after passing through the current road segment unit;

[0119] Uncertainty parameter acquisition module 55 is used to obtain the uncertainty parameter of the expected arrival time of the link in the next road segment unit based on the obtained expected arrival time of the current link, the obtained uncertainty parameter of the expected arrival time of the current link, and the feature set of the current road segment unit.

[0120] The estimated travel time acquisition module 51 is also used to take the link in the next road segment unit as the current link.

[0121] In one embodiment, the estimated travel time acquisition module 51 and the uncertainty parameter determination module 55 are used to obtain the estimated travel time and uncertainty parameters respectively in the following manner: the estimated arrival time of the current link, the uncertainty parameter of the estimated entry time of the current link, and the feature set of the current road segment unit are input into a pre-trained recurrent neural network model, and the recurrent neural network model outputs the estimated travel time of the current road segment unit and the uncertainty parameter of the estimated entry time of the current link in the current road segment unit.

[0122] In one embodiment, the feature set of the current road segment unit includes: a discrete feature set and a continuous feature set;

[0123] The discrete feature set includes at least: the road type feature of the link in the road segment unit, the road congestion status feature of the link in the road segment unit, the type feature of the intersection in the road segment unit, and the navigation action feature of the route at the intersection in the road segment unit;

[0124] The continuous feature set includes at least: at least one historical average travel time of a link in the road segment unit, and the real-time travel time of a link in the road segment unit; the at least one historical average travel time includes: each historical average travel time corresponding to multiple historical arrival times before and / or after the estimated arrival time of the current link.

[0125] In one embodiment, the above-described apparatus for determining the estimated arrival time of a route refers to... Figure 4 As shown, it further includes: a feature processing module 56, used to input various features in the discrete feature set into the embedding layer, and use the output of the embedding layer as the input of the recurrent neural network model.

[0126] This disclosure also provides a computer-readable storage medium storing computer instructions that, when executed by a processor, implement the method described above for determining the estimated arrival time of a route.

[0127] This disclosure also provides a navigation server, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described above for determining the estimated arrival time of a route.

[0128] The method, related apparatus, and server for determining the estimated time of arrival (ETA) of a route provided in this disclosure more closely approximate the actual situation during real-world travel, where the uncertainty of the estimated ETA of the previous link affects the estimated travel time of adjacent links. This results in a more accurate estimated travel time for each road segment unit, and consequently, a more accurate estimated ETA for the entire planned navigation route. According to the inventors' calculations, in the Beijing area, the method for determining the estimated ETA using this disclosure improves the accuracy of ETA calculation by more than 3.15% compared to using a traditional baseline model, achieving superior technical results.

[0129] Furthermore, this embodiment employs a pre-trained neural network model to participate in the calculation of the estimated travel time for each road segment unit in the route and the learning of the estimated arrival time uncertainty parameter of the link. Since the neural network model can learn from real user travel data samples to obtain the weight relationship between the historical average travel times corresponding to multiple times before and after entering a link at different times, the accuracy of ETA calculation can be further improved through the learning of the neural network.

[0130] Those skilled in the art will understand that embodiments of this disclosure can be provided as methods, systems, or computer program products. Therefore, this disclosure can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this disclosure can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0131] This disclosure is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0132] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0133] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0134] Obviously, those skilled in the art can make various modifications and variations to this disclosure without departing from its spirit and scope. Therefore, if such modifications and variations fall within the scope of the claims of this disclosure and their equivalents, this disclosure is also intended to include such modifications and variations.

Claims

1. A method for determining the estimated arrival time of a route, characterized in that, The route includes at least one road segment unit, and the road segment unit includes a road link or a road link and its entry point. The following steps are performed on the road segment unit: Based on the obtained estimated arrival time of the current road segment link in the current road segment unit, the obtained uncertainty parameter of the estimated arrival time of the current road segment link, and the feature set of the current road segment unit, at least the estimated travel time of the current road segment unit is obtained; wherein, the uncertainty parameter is used to characterize the time deviation uncertainty value of the estimated arrival time of the road segment link in the current road segment unit. Obtain the sum of the estimated arrival time of the current road segment link and the estimated travel time of the current road segment unit; If the current segment unit is the last segment unit of the route, then the sum is determined as the estimated arrival time of the route.

2. The method as described in claim 1, characterized in that, If the current road segment unit is not the last road segment unit of the route, the method further includes: The sum is determined as the estimated arrival time of the link in the next segment unit, where the next segment unit is the segment unit that the route enters after passing the current segment unit. Based on the estimated arrival time of the current road segment link, the uncertainty parameter of the estimated arrival time of the current road segment link, and the feature set of the current road segment unit, the uncertainty parameter of the estimated arrival time of the road segment link in the next road segment unit is obtained. Take the link in the next segment unit as the current segment link and return the step of obtaining at least the estimated travel time of the current segment link.

3. The method as described in claim 2, characterized in that, Based on the obtained estimated arrival time of the current road segment link, the obtained uncertainty parameters of the estimated arrival time of the current road segment link, and the feature set of the current road segment unit, the uncertainty parameters of the estimated arrival time of the road segment link in the next road segment unit are obtained, including: The estimated arrival time of the current road segment link, the uncertainty parameter of the estimated arrival time of the current road segment link, and the feature set of the current road segment unit are input into a pre-trained recurrent neural network model. The recurrent neural network model outputs the estimated passage time of the current road segment unit and the uncertainty parameter of the estimated arrival time of the road segment link in the next road segment unit.

4. The method as described in claim 3, characterized in that, The feature set of the road segment unit includes: a discrete feature set and a continuous feature set; The discrete feature set includes at least: road type features of road links in the road segment unit, road congestion status features of road links in the road segment unit, type features of entry into intersections in the road segment unit, and navigation action features of the route at intersections in the road segment unit; The continuous feature set includes at least: at least one historical average travel time of a road link in the road segment unit, and the real-time travel time of a road link in the road segment unit; the at least one historical average travel time includes: each historical average travel time corresponding to multiple historical arrival times before and / or after the estimated arrival time of the current road link.

5. The method as described in claim 4, characterized in that, The method further includes: The features in the discrete feature set are input into the embedding layer, and the output of the embedding layer is used as the input of the recurrent neural network model.

6. An apparatus for determining the estimated arrival time of a route, characterized in that, include: The estimated travel time acquisition module is used to obtain at least the estimated travel time of the current road segment unit based on the estimated arrival time of the current road segment link in the current road segment unit, the uncertainty parameter of the estimated arrival time of the current road segment link, and the feature set of the current road segment unit; the route includes at least one road segment unit, and the road segment unit includes a road segment link or a road segment link and its entry intersection; wherein, the uncertainty parameter is used to characterize the time deviation uncertainty value of the estimated arrival time of the road segment link in the current road segment unit; The time accumulation module is used to obtain the sum of the estimated arrival time of the current road segment link and the estimated travel time of the current road segment unit; The judgment module is used to determine whether the current road segment link is the road segment link of the last road segment unit of the route; An estimated arrival time determination module is used to determine the estimated arrival time of the route by summing the output of the time accumulation unit when the determination module determines that the link is the last link unit of the route.

7. A computer-readable storage medium storing computer instructions thereon, characterized in that, The instruction is executed by the processor as a method for determining the estimated arrival time of a route as described in any one of claims 1-5.

8. A navigation server, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method for determining the estimated arrival time of a route as described in any one of claims 1-5.

Images