Computer-implemented methods, computer programs and data processing systems for determining a travel duration

Abstract

The present disclosure relates to the general field of transportation. More specifically, this disclosure proposes computer-implemented methods of determining a travel duration for a transportation, or travel, from a departure location to a destination location. In one aspect described herein, a departure location and a destination location for the transportation are determined. Furthermore, an estimation of a driving time from the determined departure location to the determined destination location is produced. An input data set related to the transportation is also obtained. This input data set includes at least identification data of a courier for the transportation. A correction time estimation is produced, or created, based on the obtained input data set. This correction time estimation is specific to the courier being associated with said identification data. Furthermore, the correction time is applied to the estimation of the driving time to produce the estimation of the travel duration..

Description

[0001] COMPUTER-IMPLEMENTED METHODS, COMPUTER PROGRAMS AND DATA PROCESSING SYSTEMS FOR DETERMINING A TRAVEL DURATION

[0002] Technical Field

[0003] The present disclosure relates to the general field of transportation. More specifically, this disclosure proposes computer-implemented methods of determining a travel duration for transportation, or travel, from a departure location to a destination location. The disclosure also presents computer-implemented methods of optimizing a route between a departure location and a destination location. Corresponding computer programs, carriers and data processing systems are also described.

[0004] Background

[0005] Courier services delivering or collecting goods use collection or delivery addresses to plan courier journeys involving multiple destinations as optimally as possible, e.g. with as little expenditure of time or materials as possible.

[0006] Figs. 1A-1C illustrates an example of a common existing approach of estimating the travel duration, or journey duration, from a departure location to a destination location. In the example, the departure location is the location, or point, number 54. In the illustrative map, departure location #54 can be found a bit north of Holmenkollen. In addition, the destination location is assumed to be location #55, which can be found farther northwest of Holmenkollen. As can be seen in more detail in Fig. IB, a typical travel duration from departure location #54 to destination location #55 includes a driving time and service time. In the common existing approach of estimating 100 the travel duration, the procedure generally involves estimating 110 the driving time and setting 120 the service time. This approach is sometimes referred to as the Common Approach (CA) throughout this disclosure.

[0007] Other approaches exist too. For example, in an advanced common approach (sometimes referred to herein as the Advanced Common Approach (AC A)) the calculated driving time itself is adjusted by a certain estimation. In ACA, there are the calculation of the driving time, the estimation of the adjustment of the driving time and the estimation of the service time. These three calculations / estimations are relatively complex and may thus drive costs.

[0008] Summary

[0009] It is in view of the above considerations and others that the various embodiments of the present invention have been made.

[0010] The present disclosure recognizes the fact that the common existing approach discussed in connection with Figs. 1 A-1C for estimating a travel duration may be inadequate.

[0011] The travel duration from a point A (departure) to a point B (destination) typically includes a driving time and a service time as can be seen in Figs. 1 A and IB. The driving time is the time it takes to travel between A and B, here exemplified by deliveries #54 and #55. The service time is the time it takes for the courier to perform a service at point B (here exemplified by the delivery #55). The service time may be affected by various factors inter alia.

[0012] - Nature of goods'. The type of goods being delivered or collected can significantly impact the service time. Fragile or perishable items may require extra care and handling, leading to longer service times. Package size and weight'. Larger or heavier packages may take more time to handle, load, and unload. Couriers may need additional time to manage oversized or bulky items.

[0013] Special handling requirements'. Some packages may have special handling instructions, such as "fragile," "handle with care," or "this side up." Couriers may need to take extra precautions, slowing down the service time.

[0014] - Access to delivery / pickup points'. Difficulties in accessing delivery or pickup points, such as gated communities, restricted areas, or locations without proper loading / unloading facilities, can contribute to longer service times. Additionally, or alternatively, the service time may include parking time for the courier. If so, the parking lots may be located at varying distances from the destination itself, which could extend the walking time needed for the courier to access the delivery / pickup point.

[0015] For example, this approach, i.e. the common existing approach discussed in connection with Figs. 1 A-1C, generally suffers from various sources of error. Errors may for instance come from incorrect, or inaccurate, estimations of the driving time. Additionally, or alternatively, errors may come from a wrongly set service time. The total sum of potential errors may become substantial in certain scenarios.

[0016] The present disclosure therefore recognizes the fact that there is a need for alternatives to (e.g. improvement of) the existing art described above. It is an object of some aspects and embodiments described herein to solve, mitigate, alleviate, or eliminate one or more of the disadvantages with the common existing approach described above.

[0017] A general object of the present disclosure is therefore to provide a new approach, which is an alternative to (or, improvement over) the existing prior art. More specifically, an object of some aspects and embodiments discussed herein is to provide an improved method and system for estimating a travel duration from a departure location to a destination location. Moreover, an object of some aspects and embodiments discussed herein is to subsequently utilize the improved travel duration estimations for optimizing routes between locations. A technical purpose of providing the methods and systems proposed herein is to contribute to a reduced environmental impact caused by, for example, courier or other transport services. With improved travel duration estimations and thereby optimized routes, the vehicles used for the travels may e.g. travel shorter distances, may choose smarter routes when travelling between departure and destination locations, and / or may be able to deliver more parcels within a given time period. As will be appreciated, since more parcels may be delivered within a given time period, fewer vehicles are typically required for a same amount of parcels to be delivered. For non-electrical vehicles, e.g. vehicles using combusting engines, this also means that the CO2 footprint may be reduced. For electrical vehicles (EVs), this means that the vehicles may be charged less frequently and, in turn, this contributes to a better usage of energy resources. Still further, with improved and more reliable travel duration estimations the total number of vehicles that is needed, e.g. for courier or delivery services, can be reduced.

[0018] The above-mentioned general object has therefore been addressed by the appended independent claims. Advantageous embodiments are defined in the appended dependent claims.

[0019] According to a first aspect, a computer-implemented method of determining an estimation of a travel duration for a transportation from a departure location to a destination location is proposed. The travel duration may include a driving time and a service time.

[0020] The method comprises: determining said departure location and said destination location for the transportation; producing an estimation of a driving time from the determined departure location to the determined destination location; obtaining an input data set related to the transportation, the input data set including at least identification data of a courier for the transportation; producing a correction time estimation based on the obtained input data set, the correction time estimation being specific to the courier being associated with said identification data; and applying the correction time to the estimation of the driving time to produce the estimation of the travel duration. In some embodiments, the method may comprise determining a Time- of-Delivery (ToD) estimation and adding the estimation of the travel duration to the ToD to produce a Time-of-Next-Delivery (ToND) estimation.

[0021] In some embodiments, the input data set may further include at least one (i.e., one or more) of the following input data: data related to the determined departure location and / or data related to the determined destination location; route data (e.g., historical route data) of one or multiple transportations associated with said identification data; data related to the specific means of transportation between the departure location and the destination location (e.g., a transportation vehicle such as a motor vehicle (e.g., car, truck), a boat or a bicycle); data related to the specific time of the day and / or the specific time of the year for the transportation; - transportation requirements data; data including number and / or size of delivery items to be delivered; data related to a sequence number of a delivery (e.g., the number of the delivery in a delivery sequence of a route);

[0022] - time data; and

[0023] - weather data.

[0024] In some embodiments, producing the correction time based on the obtained input data set may include training a regression machine learning (ML) algorithm to model the correction time. In some implementations, producing the correction time estimation based on the obtained input data set may include training a regression ML algorithm to model the correction time by: initialising the regression ML algorithm; observing a first state of said input data set; observing a change from the first state to a second state, the change resulting from performance of at least one action that has an effect on at least one of the input data in the input data set; and updating the regression machine learning algorithm based on the observed change of state.

[0025] In some embodiments, the regression ML algorithm may utilize gradient boosting. For example, gradient boosting may include gradient tree boosting.

[0026] According to a second aspect, there is provided a computer program comprising instructions which, when executed on at least one processor, cause the at least one processor to carry out the method according to the first aspect.

[0027] According to a third aspect, there is provided a carrier comprising the computer program according to the second aspect. The carrier may, for example, be any one of an electronic signal, an optical signal, a radio signal, or a computer-readable medium.

[0028] According to a fourth aspect, a data processing system is provided. The data processing system may comprise at least one processor and at least one memory. The at least one memory may comprise instructions executable by the at least one processor whereby the data processing system is operative to perform the method according to the first aspect.

[0029] In some embodiments, the data processing system may include one or more servers. Advantageously, but not necessarily, several servers may be used in a distributed computing environment. The present disclosure recognizes the fact that aspects described hitherto may allow for a usage of fewer data points or fewer computing repetitions compared to the common existing approach described in the background. For example, the proposed method typically requires only the estimation of the driving time once and a single correction whereas the CA and the ACA described in the background may in some scenarios need repeated computing estimations (cf. Fig. 1C). Using fewer data points or less computing repetitions may have the further advantage of contributing to fewer possible sources of errors and fewer errors leads to improved travel duration estimations. For example, the proposed method uses only one correction and may therefor only have one error. In contrast, in the CA and the ACA it is possible that two or more corrections lead to corresponding two or more errors. Fewer errors contribute to the computational efficiency of the proposed method and also has the advantage of contributing to increasingly accurate estimations.

[0030] As will appreciated from the above discussion, the proposed method may contribute to a more computationally efficient method for arriving at as accurate travel duration estimations as possible. This computationally efficiency may reduce the computational effort of a data processing system implementing the method. Reducing the computational effort is also advantageous as this reduces the need for energy used by the data processing system. Aspects described hitherto may thus contribute to energy savings, as compared to e.g. the common existing approach.

[0031] Still further, the aspects described hitherto may have additional advantages. For example, while the traditional approach described in the background sometimes require or otherwise depends on relatively precise driving times the aspects described hitherto may instead accept lower quality driving times, i.e. driving times that are less precise. In other words, the aspects described hitherto are operable to handle low quality driving times well, because these aspects were developed based on the explicit intention not only to determine a precise service time but to correct the driving time to be precise. Since the aspect described hitherto may accept less precise driving times, it becomes possible to use or otherwise utilize inexpensive services for the provision of the data related to the driving times. Compared to some existing approaches that rely on data related to high- quality driving times which is known to be expensive, the aspects described hitherto can contribute to cost savings, potentially substantial cost savings.

[0032] According to a fifth aspect, there is provided a computer-implemented method for optimizing a route between a departure location and a destination location. The method comprises: determining a plurality of available candidate routes for transportation from said departure location to said destination location; determining an estimation of the travel duration for each one of said plurality of available candidate routes according to the first aspect described above; and selecting one of the plurality of available candidate routes to represent a most suitable route based on the determined estimations of travel duration.

[0033] In some embodiments, selecting one of the plurality of available candidate routes to represent a most suitable route may, e.g., comprise selecting the route having the shortest travel duration to represent the most suitable route.

[0034] In some embodiments, the method may further comprise communicating the selected route to a communication device associated with, or integrated into, a transport vehicle for subsequent route selection from the departure location to the destination location accordingly. For example, communicating the selected route to the communication device may comprise transmitting a data message including data related to, or otherwise indicative of, the selected route. The communication device may e.g. be a user device (e.g., a smartphone, a tablet computer or other handheld device) carried by a user and being associated with the transport vehicle in question. Alternatively, the communication device may be integrated into the transport vehicle, e.g. embodied in the navigation system or infotainment system of the transport vehicle.

[0035] According to a sixth aspect, there is provided a computer program comprising instructions which, when executed on at least one processor, cause the at least one processor to carry out the method according to the fifth aspect.

[0036] According to a seventh aspect, there is provided a carrier comprising the computer program according to the sixth aspect. The carrier may, for example, be any one of an electronic signal, an optical signal, a radio signal, or a computer-readable medium.

[0037] According to an eighth aspect, a data processing system is provided. The data processing system may comprise at least one processor and at least one memory. The at least one memory may comprise instructions executable by the at least one processor whereby the data processing system is operative to perform the method according to the first aspect.

[0038] In some embodiments, the data processing system may include one or more servers. Advantageously, but not necessarily, several servers may be used in a distributed computing environment.

[0039] The present disclosure recognizes the fact that it may be possible to optimize, or otherwise improve, the time estimation from the departure location to the destination location and optionally utilize this improved time estimation in subsequent route optimization methods. In turn, this may have the advantage that fewer computing repetitions are required compared to the CA and the ACA. This may contribute to a more computationally efficient method for optimizing a route between a departure location and a destination location. This computationally efficiency may reduce the computational effort of a data processing system implementing the method. Reducing the computational effort is also advantageous as this reduces the need for energy used by the data processing system. Aspects described hitherto may thus contribute to energy savings.

[0040] Furthermore, when courier or other transport services implement the aspects described herein they may be provided with improved travel duration estimations and possibly optimized routes. Hence, the vehicles used for the travels may use the communicated travel routes to, e.g., travel shorter distances or to choose smarter routes when travelling between departure and destination locations. For non-electrical vehicles, e.g. vehicles using combusting engines, this means that the CO2 footprint may be reduced. For electrical vehicles (EVs), this means that the vehicles may be charged less frequently and, in turn, this contributes to a better usage of energy resources. In addition, improved travel duration estimations may lead to a reduced demand on the number of transport vehicles needed to operate a specific courier or transport service.

[0041] Brief Description of the Drawings

[0042] These and other aspects, features and advantages will be apparent and elucidated from the following description of various embodiments, reference being made to the accompanying drawings, in which: Figs 1 A-1C illustrate a common approach for travel duration estimation in the existing art;

[0043] Fig. 2 is a flow chart illustrating a computer-implemented method for determining an estimation of a travel duration for a transportation from a departure location to a destination location;

[0044] Fig. 3 is a diagram illustrating a possible delivery time estimation via correction time when implementing the method shown in Fig. 2;

[0045] Fig. 4 is a flow chart illustrating a method for training a reinforcement learning algorithm to model the correction time used in the method of Fig. 2;

[0046] Fig. 5 is a cost matrix illustrating an estimation of driving times from determined departure locations to determined destination locations;

[0047] Figs 6 and 7 are cost matrices being specific to two different couriers and illustrating respective travel duration estimations;

[0048] Fig. 8 is a flow chart illustrating a computer-implemented method for optimizing a route between a departure location and a destination location;

[0049] Fig. 9 illustrates an example implementation of an embodiment of data processing system; and

[0050] Fig. 10 illustrates a carrier containing a computer program, in accordance with an embodiment.

[0051] Detailed Description of Embodiments

[0052] The present invention will now be described more fully hereinafter. The invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided by way of example so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those persons skilled in the art. Like reference numbers refer to like elements throughout the description.

[0053] As described above, the common existing approaches (a.k.a. the CA and the ACA) described in the background may be inadequate. To address this, and in accordance with an aspect of this disclosure, a computer-implemented method of determining an estimation of a travel duration for a transportation from a departure location to a destination location is proposed. This method will now be described in further detail with respect to Fig. 2 and Fig. 3

[0054] Reference is now made to Fig. 2, which shows a flowchart of the computer- implemented method 200 of determining an estimation of a travel duration for a transportation from a departure location to a destination location in accordance with an aspect. Sometimes, the expression travel duration could interchangeably be referred to as journey duration or, alternatively, delivery duration depending on the specific implementation of the aspects and embodiments described in this disclosure.

[0055] It is worth emphasizing that in the common approaches (CA and AC A) described in the background, the arrival time at the destination is generally defined as the time of the departure of the previous destination plus driving time. The time of departure at the new destination is generally defined as the time of arrival at the new destination plus service time. In the approach proposed in the following, the time of arrival and the time of departure do not exist as such. Rather, in the proposed approach only a single delivery time is defined, or otherwise utilized. In other words, the delivery time may be defined as the delivery time at the previous destination plus travel duration. The travel duration may be defined as driving time plus correction time. Here the correction time can be positive as well as negative. The proposed approach therefore does not estimate how long the driving time will be per se. Furthermore, the proposed approach does not estimate how long the service time will be per se. As will be appreciated, the proposed approach instead estimates the travel duration. The proposed approach is advantageous in that it estimates the time from the previous point of delivery to the new point of delivery as is shown in Fig. 3. This relatively simple, or non-complex, approach allows for many advantages as there are fewer error sources and thus the estimation can be made more precise. As will be appreciated, the point of delivery can be defined in a variety of ways. For example, it can be the point when the delivery is finished or if the courier starts moving again.

[0056] Action 210 The departure location is determined, or otherwise obtained. Also, the destination location is determined, or otherwise obtained. In some embodiments, the departure location and / or destination location may be manually input by the courier for the transportation, i.e. the driver who is intended to drive the transport vehicle from the departure location to the destination location. For example, the departure location and / or destination location may be manually input via a communication device associated with the courier, e.g. via a user interface of the communication device. Alternatively, departure location and / or destination location may be communicated to the communication device. The communication device may e.g. be a user device (e.g., a smartphone, a tablet computer or other handheld device) carried by a user and being associated with the transport vehicle in question. Alternatively, the communication device may be integrated into the transport vehicle, e.g. embodied in the navigation system or infotainment system of the transport vehicle.

[0057] Action 220: An estimation of a driving time (sometimes referred to as travelling time) from the determined departure location to the determined destination location is produced.

[0058] Action 230: An input data set related to the transportation is obtained.

[0059] The input data set includes at least identification data of a courier for the transportation. Said identification data of the courier can thus be seen as a specific identification, or identifier, of the courier. As such, the identification data of the courier may be referred to as Courier ID in this disclosure.

[0060] In advantageous embodiments, the input data set further includes at least one of the following input data: data related to the determined departure location and / or data related to the determined destination location; route data (e.g. historical route data) of one or multiple transportations associated with said identification data; data related to the specific means of transportation between the departure location and the destination location (e.g., transportation vehicle such as a motor vehicle (car, truck), a boat or a bicycle); data related to the specific time of the day and / or the specific time of the year for the transportation;

[0061] - transportation requirements data; data including number and / or size of delivery items to be delivered; data related to a sequence number of a delivery (e.g., the number of the delivery in a delivery sequence of a route);

[0062] - time data; and

[0063] - weather data.

[0064] Accordingly, one or more input data may make up the input data set.

[0065] As is known in this art and as will be appreciated by those skilled in this art, the input data may be obtained in various ways, e.g., depending on the nature of the input data. For instance, some input data (e.g. time data, transportation requirements data, data including number and / or size of delivery items to be delivered) may be manually input, e.g., by a courier via a user interface of the communication device. Additionally, or alternatively, some input data (e.g., whether data, data including number and / or size of delivery items to be delivered, transportation requirements data, route data of one or multiple transportations associated with said identification data) may be received from remote services (i.e., from remote servers or computers of such services). When received from remote services, the input data may for example be received in data messages communicated from said remote services.

[0066] Action 240: A correction time estimation based on the obtained input data set is produced, or otherwise determined. This correction time estimation is specific to the courier being associated with said identification data (Courier ID).

[0067] Action 250: The correction time is applied to the estimation of the driving time to produce the estimation of the travel duration.

[0068] In some embodiments, it is optional to produce a Time-of-Arrival estimation. Hence, actions 260 and 270 may optionally be added in method 200:

[0069] Action 260: For example, a Time-of-Delivery (ToD) estimation may be determined, or otherwise generated. That is, a ToD estimation of a first delivery is made, or otherwise performed.

[0070] Action 270: Once a ToD estimation has been determined in action 260, the estimation of the travel duration (action 250) may be added or otherwise appended to the ToD to produce a Time-of-Next-Delivery (ToND) estimation. In other words, it is possible to determine the travel duration to a subsequent (“next”) delivery, referred to as Delivery #2 in Fig. 3, by appending the travel duration to the ToD. It should be appreciated that the actions, or method steps, 210 through 270 do not necessarily have to be executed in the exact order as described above. The order is for illustrative purpose only and a person skilled in this art will appreciate that the actions, or method steps, may be executed in a different order. As a mere example, the action 230 of obtaining the input data set related to the transportation may in some embodiments be performed before action 220 of producing the estimation of the driving time. Similarly, action 260 does not have to be performed after action 250. In some implementations, certain actions may also be performed in parallel, or substantially in parallel.

[0071] Reference is now made to Fig. 4, which illustrates a method for modelling the correction time described in conjunction with Figs. 2 and 3. The present disclosure recognizes the fact that it may be advantageous to model the correction time using machine learning (ML). In advantageous embodiments, it is therefore proposed to produce the correction time estimation utilizing ML. For example, the action 240 of producing the correction time estimation based on the obtained input data set may therefore include training a regression ML algorithm to model the correction time. This will be discussed in the following.

[0072] Action 410: The regression ML algorithm is initialized. There is a wide variety of regression ML algorithms. A person skilled in the art will appreciate that various regression machine learning algorithm are conceivable for reducing the aspects and embodiments described herein into practice. For example, a non-linear regression ML algorithm may be utilized. In some implementations, the regression ML algorithm may utilize gradient boosting. The gradient boosting may e.g. include gradient tree boosting.

[0073] Action 420: A first state of the input data set mentioned earlier is observed.

[0074] Action 430: Subsequently, a change from the first state to a second state is observed. The change results from performance of at least one action that has an effect on at least one of the input data in the input data set.

[0075] Action 440: The regression machine learning algorithm can then be updated based on the observed change of state.

[0076] The regression ML algorithm used in the method of Fig. 4 may be represented by the following formula, or equation: f(X)=Y (1) where X is a combination of all the factors to take into account (i.e., the input data set), and Y is the predicted value (i.e., the correction time estimation).

[0077] In the following discussion, a possible implementation example according to some aspects and embodiments will be described in further detail. In this implementation example, a regression ML algorithm may be utilized for producing, or otherwise creating, the correction time estimation.

[0078] Table 1 illustrates an example input data set, e.g. a feature set of historical data. For example, the input data set may include one or more of the following in this example implementation:

[0079] Courier ID

[0080] - Location of last delivery

[0081] - Location, Delivery requirements (e.g., ID check, temperature control)

[0082] - Delivery properties (e.g., number of parcels, size)

[0083] Season, day, time.

[0084] Part 1 Train regression ML algorithm based on historical data: Given the Courier ID and other factors (here exemplified by an input data set including Latitude, Longitude, Parcel Size and Historical correction times) it is possible to learn what will be the correction time. In other words, given a combination of the Courier ID and associated additional input data it is made possible to provide, or otherwise produce, a correction time estimation. Table 1. Example input data

[0085] Part 2 Apply regression algorithm on new data. In an example as shown in Table 2, the courier is initially unknown. Therefore, the correction time is also unknown initially, i.e. before the courier has concluded his or her first deliveries, or transportations.

[0086] Table 2. No courier assigned yet.

[0087] Part 2 Apply regression ML algorithm on new data. In an example as shown in Table 3, a courier with Courier ID 123 has been assigned to a delivery, or transportation. When applying the regression ML algorithm to the combination of the Courier ID and associated additional (historical) input data it is made possible to provide, or otherwise produce, a correction time estimation. In other words, it is possible to predict the correction time. The predicted correction time will be specific to the courier being associated with, or otherwise linked to, the Courier ID #123. Since the correction time estimation, or prediction, is based on historical data associated with the Courier ID, the correction time estimation will take account of factors that are relevant to this particular courier, e.g. driving time, parking time, speed of loading / unloading deliveries (e.g., parcels), etc.

[0088] Table 3. Courier with Courier ID #123 assigned to delivery.

[0089] As will be appreciated, different couriers have different transportation speed and / or different service style. Additionally, or alternatively, different couriers have different driving style. For example, some couriers may typically park quicker than others (thus, correcting driving time) and some couriers may typically have better knowledge (than others) about buildings and therefore find the delivery destinations quicker (thus, correcting service time). Additionally, or alternatively, some couriers fill or stock the lockers quicker / slower at the destination locations (thus, correcting service time). Also, the service time may additionally or alternatively be affected by the destination location per se (e.g. home destination vs. destination at shopping mall). Table 4 exemplifies an example where a different courier with Courier ID #234 has been assigned to a delivery, or transportation. When applying the regression ML algorithm to the combination of the Courier ID and associated additional (historical) input data it is made possible to provide, or otherwise produce, a correction time estimation for this courier. As can be seen, this correction time estimation differs from that of the courier with Courier ID 123 shown in Table 3.

[0090] Table 4. Courier with Courier ID #234 assigned to delivery.

[0091] Reference is now made to Fig. 5-7, which illustrate cost matrices for a transportation from a determined departure location to a determined destination location. Fig. 5 is a general cost matrix illustrating an estimation of driving times (e.g., in hours, minutes or seconds) from determined departure locations to determined destination locations (cf. action 220 in Fig. 2).

[0092] Fig. 6 illustrates an example cost matrix which is specific to the courier being associated with Courier ID #123 including the correction time estimation that is specific to the courier associated with Courier ID #123. In a similar manner, Fig. 7 illustrates an example cost matrix which is specific to the courier being associated Courier ID #234 including the correction time estimation that is specific to the courier associated with Courier ID #234.

[0093] It will be appreciated by those skilled in this art that each courier may be associated with identification data (Courier ID) of that courier. The Courier ID is a specific identification, or identifier, of the courier in question. The present disclosure recognizes the fact that data on past deliveries, or transportations, made by different couriers may be collected and form the input data set. This input data set, which is thus associated with (or otherwise linked to) each unique Courier ID may be utilized to initialize, update (train) and apply the regression ML algorithm to produce correction time estimations, or predictions.

[0094] Compared to the existing common approaches (CA and ACA) discussed in the background, the aspects and embodiments described herein are advantageous in that the travel duration (e.g. a delivery duration) from a determined departure location to a determined destination location is estimated for each courier. Hence, the estimation of the travel duration may be tailored for each courier. As will be appreciated, this estimation of the travel duration may reflect individual characteristics of each individual courier (e.g., experience, driving styles, etc.). This has the advantage that the estimation of the travel duration can improve over approaches known in the prior art. In addition, the present disclosure recognizes the fact that the aspects and embodiments disclosed herein may use fewer data points during the training (updating) of the regression ML algorithm as discussed earlier herein. Using fewer data points may have the further advantage of contributing to fewer possible sources of errors and fewer errors may lead to improved travel duration estimations. In turn, this may contribute to a more computationally efficient method for arriving at as accurate travel duration estimations as possible. This computationally efficiency may reduce the computational effort of a data processing system implementing aspects and embodiments described herein. Reducing the computational effort is also advantageous as this reduces the need for energy used by the data processing system. Reducing the computational efficiency further enables the use of less expensive processors for the data processing. Hence, aspects and embodiments described herein contribute to energy savings, as compared to e.g. the common existing approach discussed in the background section.

[0095] Reference is now made to Fig. 8, which illustrates a possible computer- implemented method for optimizing a route between a departure location and a destination location. It should be appreciated that a route is a travel way, or travel course, taken from a starting point (departure location) to a destination point (destination location).

[0096] Action 810 Initially, a plurality of available candidate routes for transportation from said departure location to said destination location may be determined.

[0097] Action 820 An estimation of the travel duration for each one of said plurality of available candidate routes is determined. This estimation is performed in accordance with the aspects and embodiments described hitherto, i.e. with respect to Figs. 2-7. As will be appreciated, each one of the available candidate routes may therefore also take into account respective couriers, since the estimations of travel duration are specific to the respective couriers.

[0098] Action 830: One of the plurality of available candidate routes is selected to represent a most suitable route based on the determined estimations of travel duration. For example, in some embodiments, this selection may involve selecting the route having the shortest travel duration to represent the most suitable route. Additionally, or alternatively, this selection may involve selecting the route that contributes to the maximum amount of parcels being delivered in the shortest possible time. The process of selecting the most suitable route may be performed in accordance with existing optimization processes, which are known in the art. For example, it is conceivable to produce, or otherwise arrange, a matrix of all possible combinations (i.e., candidate routes) as is shown the left-hand side of Fig. 8 and to find (select) the most optimal combination to minimize the total sum. Action 840: In some embodiments, the method may also comprise communicating the selected route to a communication device.

[0099] The communication device may be associated with a transport vehicle for the subsequent route selection from the departure location to the destination location accordingly. The communication device may advantageously, but not necessarily, be a mobile communication device such as a laptop computer, a tablet computer, a mobile telephone, a cellular telephone, a smart phone or any other handheld devices.

[0100] Alternatively, the communication device may be integrated into a transport vehicle for the subsequent route selection from the departure location to the destination location accordingly. For example, the communication device may be integral with a navigation system or an infotainment system of the transport vehicle.

[0101] Implementing the method as described in conjunction with Fig. 8 may enable the optimization of a route between a departure location and a destination location while using, or otherwise utilizing, fewer data points compared to the common existing approaches (CA and ACA) discussed in the background section. This may contribute to a more computationally efficient method for optimizing a route between a departure location and a destination location. This computationally efficiency may reduce the computational effort of a data processing system implementing the method, especially when many available candidate routes and / or many available couriers exist. Reducing the computational effort is also advantageous as this reduces the need for energy used by the data processing system. Aspects described hitherto may thus contribute to energy savings.

[0102] Furthermore, when courier or other transport services implement the aspects and embodiments described herein they may be provided with improved travel duration estimations and in the long run also for improved optimizations of routes. Hence, the vehicles used for the travels may use the communicated travel routes to, e.g., travel shorter distances or to choose smarter routes when travelling between departure and destination locations. For non-electrical vehicles, e.g. vehicles using combusting engines, this means that the CO2 footprint may be reduced. For electrical vehicles (EVs), this means that the vehicles may be charged less frequently and, in turn, this contributes to a better usage of energy resources. Example implementations of embodiments of the communication device

[0103] Fig. 9 illustrates a data processing system 900, which is configured for performing or otherwise executing the methods according to the various aspects and embodiments discussed in this disclosure. As is schematically illustrated in Fig. 9, the data processing system 900 comprises hardware 910, 920, 930, 940. For example, the data processing system 900 may comprise one or more processors 910 and one or more memories 920. Also, a communications interface 930 may be provided in order to allow the data processing system 900 to communicate with other data processing systems and / or communication devices, e.g. via a network such as the Internet. To this end, the communications interface 930 may comprise a transmitter (Tx) and a receiver (Rx). Alternatively, the communications interface 930 may comprise a transceiver (Tx / Rx) combining both transmission and reception capabilities. The communications interface 930 may include a radio frequency (RF) interface allowing the data processing system 900 to communicate with other data processing systems and / or communication devices through a radio frequency band through the use of different radio frequency technologies such as 5G NR (New Radio), LTE (Long Term Evolution), WCDMA (Wideband Code Division Multiple Access), or any other cellular network standardized by the 3rd Generation Partnership Project (3GPP), or any other wireless technology such as Wi-Fi, Bluetooth®, etcetera. The data processing system 900 may optionally also comprise a user interface 940.

[0104] Example computer-readable mediums

[0105] Turning now to Fig. 10, another aspect will be briefly discussed. Fig. 10 shows an example of a computer-readable medium, in this example in the form of a data disc 1000. In one embodiment the data disc 1000 is a magnetic data storage disc. The data disc 1000 is configured to carry instructions 1100 that can be loaded into a memory 920 of a data processing system 900. Upon execution of said instructions by a processor 910 of the data processing system 900, the data processing system 900 is caused to execute a method or procedure according to the embodiments disclosed in this disclosure. The data disc 1000 is arranged to be connected to or within and read by a reading device (not shown), for loading the instructions into the processor. One such example of a reading device in combination with one (or several) data disc(s) 1000 is a hard drive. It should be noted that the computer-readable medium can also be other mediums such as compact discs, digital video discs, flash memories or other memory technologies commonly used. In such an embodiment the data disc 1000 is one type of a tangible computer-readable medium. The instructions may alternatively be downloaded to a computer data reading device, such as the data processing system 900 or a other computer capable of reading computer coded data on a computer-readable medium, by comprising the instructions in a computer-readable signal (not shown) which is transmitted via a wireless (or wired) interface (for example via the Internet) to the computer data reading device for loading the instructions into a processor 910 of the data processing system 900. In such an embodiment, the computer-readable signal is one type of a non-tangible computer- readable medium.

[0106] Alternative use case scenarios

[0107] Modifications and other variants of the aspects described hitherto will come to mind to one skilled in the art having benefit of the teachings presented in the foregoing description and associated drawings. Therefore, it is to be understood that this disclosure is not limited to the specific example aspects and embodiments described above. For example, a person skilled in the art will appreciate that the aspects and embodiments described herein can be useful also with means of transportations other than those specifically mentioned or listed above (i.e., typically various kinds of transport vehicles). For example, various aspects and embodiments described herein could equivalently be applied by other means of transportation, such as walking or by electrical scooters to name only a two other conceivable examples. In the following, a conceivable aspect with a drone will be described in further detail.

[0108] In one such aspect, a computer-implemented method of determining an estimation of a travel, or journey, duration for a drone transportation from a departure location to a destination location is proposed. The travel, or journey, duration may include a flight time and a service time.

[0109] The method comprises: determining said departure location and said destination location for the drone transportation; producing an estimation of a travelling time from the determined departure location to the determined destination location; obtaining an input data set related to the drone transportation, the input data set including at least identification data of the drone; producing a correction time estimation based on the obtained input data set, the correction time estimation being specific to the drone which is associated with said identification data; and applying the correction time to the estimation of the travelling time to produce the estimation of the travel duration.

[0110] In some embodiments, the method may comprise determining a ToD estimation and adding the estimation of the travel duration to the ToD to produce a ToND estimation.

[0111] In some embodiments, the input data set may further include at least one (i.e., one or more) of the following input data: data related to the determined departure location and / or data related to the determined destination location; route data (e.g., historical route data) of one or multiple drone transportations associated with said identification data; data related to the specific time of the day and / or the specific time of the year for the transportation; drone transportation requirements data; data including number and / or size of delivery items to be delivered by the drone; data related to a sequence number a delivery;

[0112] - time data; and

[0113] - weather data.

[0114] In some embodiments, producing the correction time based on the obtained input data set may include training a regression machine learning (ML) algorithm to model the correction time. In some implementations, producing the correction time estimation based on the obtained input data set may include training a regression ML algorithm to model the correction time by: initialising the regression ML algorithm; observing a first state of said input data set; observing a change from the first state to a second state, the change resulting from performance of at least one action that has an effect on at least one of the input data in the input data set; and updating the regression machine learning algorithm based on the observed change of state.

[0115] In some embodiments, the regression ML algorithm may utilize gradient boosting. For example, gradient boosting may include gradient tree boosting.

[0116] In addition, there is provided a computer-implemented method for optimizing a route between a departure location and a destination location. The method comprises: determining a plurality of available candidate routes for drone transportation from said departure location to said destination location; determining an estimation of the travel duration for each one of said plurality of available candidate routes according to the aspect with the drone described above; and selecting one of the plurality of available candidate routes to represent a most suitable route based on the determined estimations of travel duration.

[0117] In some embodiments, selecting one of the plurality of available candidate routes to represent a most suitable route may, e.g., comprise selecting the route having the shortest travel duration to represent the most suitable route.

[0118] In some embodiments, the method may further comprise communicating the selected route to a communication device associated with, or integrated into, the drone for subsequent route selection from the departure location to the destination location accordingly. For example, communicating the selected route to the communication device may comprise transmitting a data message including data related to, or otherwise indicative of, the selected route. The communication device may e.g. be a user device (e.g., a smartphone, a tablet computer or other handheld device) carried by a user and being associated with the drone in question. Alternatively, the communication device may be integrated into the drone, e.g. embodied in the drone.

[0119] Corresponding computer programs, carriers and data processing systems are also conceivable.

[0120] As will be appreciated by those skilled in the art, modifications and other variants of the described aspects and embodiments will come to mind to one skilled in the art having benefit of the teachings presented in the foregoing description and associated drawings. Therefore, it is to be understood that the embodiments are not limited to the specific example embodiments described in this disclosure and that modifications and other variants are intended to be included within the scope of this disclosure. Furthermore, although specific terms may be employed herein, they are used in a generic and descriptive sense only and not for purposes of limitation. Therefore, a person skilled in the art would recognize numerous variations to the described embodiments that would still fall within the scope of the appended claims. As used herein, the terms “comprise / comprises” or “include / includes” do not exclude the presence of other elements or steps. Furthermore, although individual features may be included in different claims, these may possibly advantageously be combined, and the inclusion of different claims does not imply that a combination of features is not feasible and / or advantageous. In addition, singular references do not exclude a plurality.

Claims

CLAIMS1. A computer-implemented method (200) of determining an estimation of a travel duration for a transportation from a departure location to a destination location, the method comprising: determining (210) said departure location and said destination location for the transportation; producing (220) an estimation of a driving time from the determined departure location to the determined destination location; obtaining (230) an input data set related to the transportation, the input data set including at least identification data of a courier for the transportation; producing (240) a correction time estimation based on the obtained input data set, the correction time estimation being specific to the courier being associated with said identification data; and applying (250) the correction time to the estimation of the driving time to produce the estimation of the travel duration.

2. The computer-implemented (200) method according to claim 1, comprising: determining (260) a Time-of-Delivery, ToD, estimation; and adding (270) the estimation of the travel duration to the ToD to produce a Time-of-Next-Delivery, ToND, estimation.

3. The computer-implemented method (200) according to claim 1 or 2, wherein the input data set further includes at least one of the following input data: data related to the determined departure location and / or data related to the determined destination location; route data of one or multiple transportations associated with said identification data; data related to the specific means of transportation between the departure location and the destination location;data related to the specific time of the day and / or the specific time of the year for the transportation;- transportation requirements data; data including number and / or size of delivery items to be delivered; data related to a sequence number of a delivery;- time data; and- weather data. The computer-implemented method (200) according to any one of the claims 1-3, wherein producing (240) the correction time estimation based on the obtained input data set includes training a regression machine learning algorithm to model the correction time by: initialising (410) the regression machine learning algorithm; observing (420) a first state of said input data set; observing (430) a change from the first state to a second state, the change resulting from performance of at least one action that has an effect on at least one of the input data in the input data set; and updating (440) the regression machine learning algorithm based on the observed change of state. The computer-implemented method (200) according to claim 4, wherein the regression machine learning algorithm utilizes gradient boosting. The computer-implemented method (200) according to claim 5, wherein gradient boosting includes gradient tree boosting. A computer program comprising instructions (1100) which, when executed on at least one processor (910), cause the at least one processor (910) to carry out the method according to any one of the claims 1-6.A carrier comprising the computer program according to claim 7, wherein the carrier is any one of an electronic signal, an optical signal, a radio signal, or a computer-readable medium (1000). A data processing system (900), comprising: at least one processor (910); and at least one memory (920), wherein the at least one memory (920) comprises instructions executable by the at least one processor (910) whereby the data processing system (900) is operative to perform the method according to any one of the claims 1-6.

Images