WO2018113974A1

WO2018113974A1 - Method and devices for predicting mobility of a mobile communication device in a cellular communication network

Info

Publication number: WO2018113974A1
Application number: PCT/EP2016/082387
Authority: WO
Inventors: Mathieu Leconte; Moez DRAIEF
Original assignee: Huawei Technologies Co., Ltd.
Priority date: 2016-12-22
Filing date: 2016-12-22
Publication date: 2018-06-28
Also published as: CN110121891B; CN110121891A

Abstract

The invention relates a method of predicting mobility of a mobile communication device (201) in a cellular communication network (200), the cellular communication network (200) comprising a plurality of network cells, including a current network cell, the method comprising: determining position information of the mobile communication device (201), wherein the position information comprises a current position L of the mobile communication device (201) in the current network cell; providing destination information of the mobile communication device (201) on the basis of the current position L of the mobile communication device (201), wherein the destination information comprises at least one likely destination D of the mobile communication device (201) and/or a destination likelihood distribution p(D) defining at least one likely destination D of the mobile communication device (201); and predicting at least a portion of a future trajectory of the mobile communication device (201) between the current position L and the likely destination D on the basis of a subset of a set of recorded local trajectories T, wherein each trajectory of the subset of the set of recorded local trajectories is associated with the likely destination D of the mobile communication device (201).

Description

DESCRIPTION

Method and devices for predicting mobility of a mobile communication device in a cellular communication network

TECHNICAL FIELD

The present invention relates to the field of wireless communications. More specifically, the invention relates to a method and system for predicting mobility of a mobile communication device in a spatial area covered by a cellular communication network.

BACKGROUND

The prediction of user mobility is a major research area in wireless networks. With the wealth of location-based services in wireless technology, a fast and accurate identification of user mobility patterns is paramount for deploying the appropriate solutions at both the network and application levels. Accurate prediction of user mobility allows for efficient planning and management of bandwidth resources and, thus, improves an abundance of services provided to mobile users, such as mobile online advertising, map adaptation, traffic information, weather forecast, smart handover management, to name a few. Moreover, mobility prediction can be used as a primitive for many optimizations in cellular

communication networks.

A number of algorithms for user mobility prediction have been reported in the literature. In general, existing approaches to mobility prediction can be cast in one of two families, namely on the one hand models based on a first approach that attempts to predict points of interest for a user - for instance, learning a map of points of interest for each user as well as associated transportation probabilities, and on the other hand models based on a second approach that attempts to predict future short-term mobility based on past short-term observations. For example, a base station A can record the mobility of a user device within its coverage area or cell and try to decide whether the user device proceeds to area B or area C, as illustrated in figure 1.

Conventional methods for user mobility prediction are still hampered by several challenging technical issues. Methods based on the first approach require collecting extensive personal data from each user, which is both subject to privacy issues and very expensive to implement. As for methods based on the second approach, past short-term observations are often not very informative to predict future mobility when users move along constrained paths (e.g., roads). Therefore, it is typically difficult to gather enough individual data to make accurate predictions due to the privacy issue, battery limitation of the user device, excessive data, etc., and the complexity of making accurate predictions is still prohibitive. Thus, in light of the above there is a need for more efficient and accurate methods and devices allowing improving the mobility prediction of a mobile communication device in a cellular communication network.

SUMMARY

It is an object of the invention to provide for more efficient and accurate methods and devices for mobility prediction of a mobile communication device in a cellular communication network.

The foregoing and other objects are achieved by the subject matter of the independent claims. Further implementation forms are apparent from the dependent claims, the description and the figures.

Generally, the present invention provides a collaborative approach between a mobile communication device, such as a mobile phone, and a communication network entity, such as a base station or an application server of a communication network, that overcome the major architectural as well as algorithmic hurdles of wireless user mobility prediction. More specifically, the present invention intelligently combines two components: a mobile communication device which maintains the user source position and estimated destinations, and a communication network entity which maintains a distribution of local trajectories together with their associated source-destination pairs. The two components can exchange the information they maintain to determine the most likely future trajectories the mobile communication device will follow and, thus, estimate its next position.

According to the present invention only very little private data is disclosed, as the private data regarding the user mobility is known only to the mobile communication device and the communication network entity can only remember aggregate data collected from multiple mobile communication devices. The mobile communication device can choose the level of accuracy it is willing to disclose. In addition, the communication network entity may only have access to local data. In other words, the present invention can also function in the absence of global data, yet keeping data local. This feature significantly reduces both the cost and the complexity of handling the data as well as computing predictions from it. Thus, according to a first aspect the invention relates a method of predicting mobility of a mobile communication device in a spatial area covered by a cellular communication network, the cellular communication network comprising a plurality of network cells, including a current network cell. The method comprises: determining position information of the mobile communication device, wherein the position information comprises a current position L of the mobile communication device in the current network cell; providing destination information of the mobile communication device on the basis of the current position L of the mobile communication device, wherein the destination information comprises at least one likely destination D of the mobile communication device and/or a destination likelihood distribution p(D) defining at least one likely destination D of the mobile communication device; and predicting at least a portion of a future trajectory of the mobile communication device between the current position L and the likely destination D on the basis of a subset of a set of recorded local trajectories T, wherein each trajectory of the subset of the set of recorded local trajectories is associated with the likely destination of the mobile communication device.

As will be appreciated, a destination likelihood distribution p(D) could define several possibly equally likely destinations D of the mobile communication device. In an implementation form, the likely destination D can be a most likely destination D. The subset of the set of recorded local trajectories T can be considered to be a filtered set of the recorded local trajectories T, i.e. from the whole set of recorded local trajectories T those are selected that are associated with the likely destination D. In an implementation form, the association between recorded local trajectories and the likely destination can be achieved by labelling each recorded local trajectory with a destination. Thus, a more efficient and accurate method is provided, allowing the improvement of predicting mobility of the mobile communication device in the cellular communication network. Memory requirements are reduced for both the mobile communication device and the cellular communication network, because the mobile communication device only needs to store its own points of interest, (i.e., the possible destinations it may visit), and in order to be able to provide a prediction for its destination the cellular communication network only needs to record the local trajectories T of the mobile communication device (as well as other mobile communication devices), but not its identity information or points of interest. A big reduction in memory and complexity comes from the trajectories being local ones only, and not global trajectories.

Moreover, privacy concerns are mitigated, as the mobile communication device can tune how accurate the locations it is willing to provide to trade its privacy off against the performance, and the cellular communication network does not need to store the identity information of the mobile communication device.

In a first implementation form of the method according to the first aspect as such, each trajectory of the subset of the set of recorded local trajectories comprises a first portion, wherein the first portion coincides at least within the current network cell with a global trajectory between the current position L and the likely destination D of the mobile communication device. In a second implementation form of the method according to the first aspect as such or the first implementation form thereof, the position information further comprises a previous position S, for instance the initial source position, of the mobile communication device and each trajectory of the subset of the set of recorded local trajectories further comprises a second portion between the previous position S and the current position L of the mobile communication device.

In a third implementation form of the method according to the first aspect as such or the first or second implementation form thereof, the position information further comprises a past trajectory H of the mobile communication device between the previous position S and the current position L and each trajectory of the subset of the set of recorded local trajectories further comprises a second portion corresponding to the past trajectory H of the mobile communication device between the previous position S and the current position L of the mobile communication device.

In a fourth implementation form of the method according to the third implementation form of the first aspect, the step of predicting at least a portion of the future trajectory T of the mobile communication device between the current position L and the likely destination D on the basis of the subset of the set of recorded local trajectories comprises the step of determining the following conditional likelihood distribution: p^*(T I H,L,S) =∑D p(D) p^*(T I H,L,S-D), wherein∑D denotes the sum over all destinations and wherein the conditional distribution p^*(T I H,L,S-D) is based on the following equation: p^*(T I H.L.S-D) = p^*(T,H,L,S-D) /∑_T p^*(T,H,L,S-D), wherein p^*(T,H,L,S-D) denotes the distribution over the set of recorded local trajectories T and∑T denotes the sum over the set of recorded local trajectories T.

In a fifth implementation form of the method according to the third or fourth implementation form of the first aspect, the past trajectory H of the mobile communication device is the past trajectory of the mobile communication device in the current network cell of the cellular communication network, i.e. the previous position S is the position where the mobile communication device has entered the current network cell. In another implementation form, the previous position S is the position, where the mobile communication device initially has connected to the cellular communication network, i.e. has attached to the cellular

communication network.

In a sixth implementation form of the method according to any one of the third to fifth implementation form of the first aspect, the method comprises the further step of refining the destination information of the mobile communication device on the basis of the past trajectory H between the previous position S and the current position L of the mobile communication device.

Refining the destination information of the mobile communication device can help to more precisely predict the future trajectory T of the mobile communication device and, thus, provide a better estimate for the most likely destination D, which can serve as a basis for other prediction services.

In a seventh implementation form of the method according to the sixth implementation form of the first aspect, the step of refining the destination information comprises the step of determining a refined destination likelihood distribution on the basis of the following equations: p^*(D I H,L,S) =∑T p^*(T,H,L,S-D) /∑_D,_T p^*(T,H,L,S-D), p(D) <- p(D) p^*(D I H,L,S) /∑_D p(D) p^*(D | H,L,S) wherein∑D denotes the sum over all destinations,∑τ denotes the sum over all recorded local trajectories and∑D,T denotes the sum over all destinations and all recorded local trajectories.

Thus, a feedback from the cellular communication network based on the past trajectory H between the previous position S and the current position L of the mobile communication device is provided, improving the prediction accuracy of the destination. In an eighth implementation form of the method according to the first aspect as such or any one of the first to seventh implementation form thereof, the future trajectory of the mobile communication device between the current position L and the most likely destination D is predicted on the basis of the subset of the set of recorded local trajectories T associated with the likely destination by selecting the trajectory of the subset of the set of recorded local trajectories between the current position L and the most likely destination D that occurs most often in the subset of the set of recorded local trajectories.

Thus, the future trajectory T of the mobile communication device can be accurately predicted with the collaborative inputs from the mobile communication device and the cellular communication network, allowing the improvement of user services which require accurate prediction of user mobility.

In a ninth implementation form of the method according to the first aspect as such or any one of the first to eighth implementation form thereof, the step of providing destination information comprises the step of selecting the likely destination D of the mobile communication device from a set of points of interest associated with the mobile communication device on the basis of the current position L of the mobile communication device and/or the step of determining the destination likelihood distribution p(D) using a set of points of interest associated with the mobile communication device on the basis of the current position L of the mobile

communication device.

In a tenth implementation form of the method according to the first aspect as such or any one of the first to ninth implementation form thereof, the method comprises the further step of providing context information associated with the mobile communication device and wherein the step of predicting at least a portion of the future trajectory of the mobile communication device comprises predicting at least a portion of the future trajectory of the mobile communication device between the current position L and the likely destination D on the basis of the subset of the set of recorded local trajectories T and on the basis of the context information. In an implementation form the context information can comprise information about a current transportation means of the mobile communication device and/or a user profile associated with the mobile communication device.

Thus, the context information of the mobile communication device will largely benefit the cellular communication network to improve the prediction accuracy of the future trajectory T of the mobile communication device. According to a second aspect the invention relates to a mobile communication device for communication in a cellular communication network, the cellular communication network comprising a plurality of network cells, including a current network cell, wherein the mobile communication device comprises: a communication interface for communication with a communication network entity of the cellular communication network; and a processor configured to: provide destination information of the mobile communication device on the basis of a current position L of the mobile communication device in the current network cell and/or a previous position S of the mobile communication device, wherein the destination information comprises a at least one likely destination D of the mobile communication device and/or a destination likelihood distribution p(D) defining at least one likely destination D of the mobile communication device; provide the destination information via the communication interface to the communication network entity; and refine the destination information of the mobile communication device on the basis of refined destination information provided by the communication network entity via the communication interface, wherein the refined destination information comprises at least a portion of a future trajectory of the mobile communication device between the current position L and the likely destination D, a refined at least one likely destination D of the mobile communication device and/or a refined destination likelihood distribution p(D) defining at least one likely destination D of the mobile communication device.

Thus, the mobile communication device can provide the communication network entity with the source information and the destination information, improving the prediction accuracy of the future mobility of the mobile communication device. According to a third aspect the invention relates to a communication network entity for predicting mobility of a mobile communication device in a cellular communication network, the cellular communication network comprising a plurality of network cells, including a current network cell, wherein the communication network entity comprises: a communication interface for receiving destination information from the mobile communication device, wherein the destination information comprises at least one likely destination D of the mobile communication device and/or a destination likelihood distribution p(D) defining at least one likely destination D of the mobile communication device; a memory configured to store a set of recorded local trajectories T; a processor configured to determine position information of the mobile communication device, wherein the position information comprises a current position L of the mobile communication device in the current network cell, and to predict at least a portion of a future trajectory of the mobile communication device between the current position L and the likely destination D on the basis of a subset of the set of recorded local trajectories T, wherein each trajectory of the subset of the set of recorded local trajectories is associated with the likely destination D of the mobile communication device.

Thus, the communication network entity can receive the source information and the destination information from the mobile communication device, improving the prediction accuracy of the future mobility of the mobile communication device. In an implementation form, the communication network entity is a base station or an application server of the cellular communication network. In a first implementation form of the communication network entity according to the third aspect as such, each trajectory of the subset of the set of recorded local trajectories comprises a first portion, wherein the first portion coincides at least within the current network cell with a global trajectory between the current position L and the likely destination D of the mobile communication device.

In a second implementation form of the communication network entity according to the third aspect as such or the first implementation form thereof, the position information further comprises a previous position S, for instance the initial source position, of the mobile communication device and each trajectory of the subset of the set of recorded local trajectories further comprises a second portion between the previous position S and the current position L of the mobile communication device.

In a third implementation form of the communication network entity according to the third aspect as such or the first or second implementation form thereof, the position information further comprises a past trajectory H of the mobile communication device between the previous position S and the current position L and each trajectory of the subset of the set of recorded local trajectories further comprises a second portion corresponding to the past trajectory H of the mobile communication device between the previous position S and the current position L of the mobile communication device.

In a fourth implementation form of the communication network entity according to the third aspect as such or any one of the first to third implementation form thereof, the processor is further configured to determine a refined estimate of the likely destination D of the mobile communication device and to feed the refined estimate of the most likely destination D via the communication interface back to the mobile communication device and/or the processor is further configured to determine a refined destination likelihood distribution p(D) and to feed the refined destination likelihood distribution p(D) back to the mobile communication device. Thus, the communication network entity can also benefit by providing the mobile communication device with the refined likely destination D and/or the refined destination likelihood distribution p(D). In a fifth implementation form of the communication network entity according to the third aspect as such or any one of the first to fourth implementation form thereof, the memory is configured to store each local trajectory of the set of recorded local trajectories T together with the most likely destination D associated therewith. In a further implementation form, the memory can store in addition to the most likely destination D associated with local trajectory, for instance, the source S and/or some other context information.

According to a fourth aspect the invention relates to a computer program comprising program code for performing the method according to the first aspect or any one of its implementation forms when executed on a computer.

The invention can be implemented in hardware and/or software.

BRIEF DESCRIPTION OF THE DRAWINGS Further embodiments of the invention will be described with respect to the following figures, wherein:

Fig. 1 shows a schematic diagram of a cellular communication network illustrating the basic concept of predicting the mobility of a mobile communication device by a base station;

Fig. 2 shows a schematic diagram of a communication network comprising a mobile communication device and a communication network entity according to an embodiment;

Fig. 3 shows a schematic diagram of the mobile communication device and the communication network entity of figure 2 in a first communication stage;

Fig. 4 shows a schematic diagram of the mobile communication device and the communication network entity of figure 2 in a second communication stage; Fig. 5 shows a schematic diagram illustrating the interaction between a mobile communication device and a communication network entity according to an embodiment for predicting mobility of the mobile communication device; Fig. 6 shows a schematic diagram illustrating a method of predicting mobility of a mobile communication device in a cellular communication network according to an embodiment;

Fig. 7 shows simulation results for a mobile communication device and a

communication network entity according to an embodiment; and

Fig. 8 shows simulation results for a mobile communication device and a

communication network entity according to an embodiment. In the various figures, identical reference signs will be used for identical or at least functionally equivalent features.

DETAILED DESCRIPTION OF EMBODIMENTS In the following description, reference is made to the accompanying drawings, which form part of the disclosure, and in which are shown, by way of illustration, specific aspects in which the present invention may be placed. It will be appreciated that other aspects may be utilized and structural or logical changes may be made without departing from the scope of the present invention. The following detailed description, therefore, is not to be taken in a limiting sense, as the scope of the present invention is defined by the appended claims.

For instance, it will be appreciated that a disclosure in connection with a described method may also hold true for a corresponding device or system configured to perform the method and vice versa. For example, if a specific method step is described, a corresponding device may include a unit to perform the described method step, even if such unit is not explicitly described or illustrated in the figures.

Moreover, in the following detailed description as well as in the claims embodiments with different functional blocks or processing units are described, which are connected with each other or exchange signals. It will be appreciated that the present invention covers embodiments as well, which include additional functional blocks or processing units that are arranged between the functional blocks or processing units of the embodiments described below. Finally, it is understood that the features of the various exemplary aspects described herein may be combined with each other, unless specifically noted otherwise.

Figure 2 shows a schematic diagram of a cellular communication network 200 comprising a mobile communication device 201 and a communication network entity 202 according to an embodiment. The cellular communication network 200 comprises a plurality of network cells, including a network cell the mobile communication device 201 is currently located in (which will be referred to herein as the current network cell). The mobile communication device 201 is configured for cellular communication with the communication network entity 202 of the cellular communication network 300.

As can be taken from the detailed view of the mobile communication device 201 shown in figure 2, the mobile communication device 201 comprises a communication interface 201 a and a processor 201 b. The processor 201 b is configured to provide destination information of the mobile communication device 201 on the basis of a current position L of the mobile communication device 201 in the current network cell to the communication network entity 202. The destination information can comprise one or more likely destinations D of the mobile communication device 201 and/or a destination likelihood distribution p(D) defining the one or more likely destination D of the mobile communication device 201 .

As will be appreciated, the destination likelihood distribution p(D) could define several possibly equally likely destinations D of the mobile communication device 201. In an embodiment, the likely destination D can be the most likely destination D of the mobile communication device 201 .

In an embodiment, the processor 201 b of the mobile communication device 201 is configured to provide the destination information by selecting the likely destination D of the mobile communication device 201 from a set of points of interest stored in the mobile

communication device 201 on the basis of the current position L of the mobile

communication device 201 or by determining the destination likelihood distribution p(D) using a set of points of interest stored in the mobile communication device 201 on the basis of the current position L of the mobile communication device 201 . The communication network entity 202 is configured to predict the mobility, i.e. a future trajectory of the mobile communication device 201 in the cellular communication network 200. In an embodiment, the communication network entity 202 is implemented as a base station or a part thereof or as an application server of the cellular communication network 200.

The communication network entity 202 comprises a communication interface 202a for receiving the destination information from the mobile communication device 201. Moreover, the communication network entity 202 comprises a memory 202c configured to store a set of recorded local trajectories T, as will be explained in more detail further below. Finally, the communication network entity 202 comprises a processor 202b configured to determine position information of the mobile communication device 201 , wherein the position information comprise a current position L of the mobile communication device 201 in the current network cell.

Moreover, the processor 202b of the communication network entity 202 is configured to predict at least a portion of a future trajectory of the mobile communication device 201 between the current position L and the likely destination D on the basis of a subset of the set of recorded local trajectories T, wherein each trajectory of the subset of the set of recorded local trajectories is associated with the likely destination D of the mobile communication device 201 . The subset of the set of recorded local trajectories T can be considered to be a filtered set of the recorded local trajectories T, i.e. from the whole set of recorded local trajectories T those are selected that are associated with the likely destination D of the mobile communication device 201. The association between recorded local trajectories and the likely destination D can be achieved beforehand, for instance, by labelling each recorded local trajectory with its destination.

Each trajectory of the set of recorded local trajectories T can define a sequence of geographical locations/positions, possibly with associated timing information. The

geographical locations/positions can be provided with different levels of accuracy, or area- based (e.g., segments of road, cells of a network).

In an embodiment, the set of recorded local trajectories T can be generated in the following manner. In a first stage, points of interests of users, i.e. locations, where the users spend a large amount of time are identified. In a second stage, trajectories are segmented such that these start or end at identified points of interest. In a third stage, the trajectories are labelled by the destination D and possibly by their source S. In a fourth stage, the trajectories are segmented once more in accordance with the area covered by the communication network entity 202 and recorded in the memory 202c of the communication network entity 202. In a further embodiment, the set of recorded local trajectories T can be generated in the following manner. Each time a new local trajectory of a mobile communication device is observed by the communication network entity 202, the trajectory is labelled with the source information S and no destination information D for the time being. When the mobile communication device reaches its destination D (which the mobile communication device itself device or a communication network entity, i.e. base station serving the current cell of the mobile communication device can identify), information about the destination D is send to the base stations along the past trajectory of the mobile communication device. These base stations can now update the destination label D as well and incorporate the local trajectory in their database of recorded local trajectories.

In a further embodiment, the set of recorded local trajectories T can be generated in the following manner. Each time a new trajectory of a mobile communication device is observed by the communication network entity 202, it is labelled with the source information S and previous destination prediction available from the mobile communication device before it leaves the cell. In this embodiment, some trajectories may be incorrectly labelled due to an inaccurate prediction of the destination D. However, this improves as the mobile

communication device proceeds towards the destination D, because estimates of the destination D get refined.

In an embodiment, each trajectory of the subset of the set of recorded local trajectories comprises a first portion, wherein the first portion coincides at least within the current network cell with a global trajectory between the current position L and the likely destination D of the mobile communication device 201.

In an embodiment, the position information further comprises a previous position S, for instance the initial source position, of the mobile communication device 201 and each trajectory of the subset of the set of recorded local trajectories further comprises a second portion between the previous position S and the current position L of the mobile

communication device 201 .

In an embodiment, the position information further comprises a past trajectory H of the mobile communication device 201 between the previous position S and the current position L and each trajectory of the subset of the set of recorded local trajectories further comprises a second portion corresponding to the past trajectory H of the mobile communication device 201 between the previous position S and the current position L of the mobile communication device 201 . In an embodiment, the processor 202b of the communication network entity 202 is configured to predict the future trajectory of the mobile communication device 201 between the current position L and the likely destination D on the basis of the subset of the set of recorded local trajectories T associated with the likely destination D by selecting the trajectory of the subset of the set of recorded local trajectories associated with the likely destination D that occurs most often in the subset of the set of recorded local trajectories associated with the likely destination D.

In an embodiment, the processor 202b of the communication network entity 202 is further configured to determine a refined estimate of the likely destination D of the mobile communication device 201 and to feed the refined estimate of the likely destination D via the communication interface 202a back to the mobile communication device 201. Additionally or alternatively, the processor 202 of the communication network entity 202 is further configured to determine a refined destination likelihood distribution p(D) and to feed the refined destination likelihood distribution p(D) via the communication interface 202a back to the mobile communication device 201.

In an embodiment, the processor 201 b of the mobile communication device 201 is configured to refine its destination information on the basis of refined destination information provided by the communication network entity 202 via the communication interface 201 a. The refined destination information comprises at least a portion of a future trajectory of the mobile communication device 201 between the current position L and the likely destination D, a refined at least one likely destination D of the mobile communication device 201 and/or a refined destination likelihood distribution p(D) defining at least one likely destination D of the mobile communication device 201 .

Further embodiments of the mobile communication device 201 and the network

communication entity 202 are shown in figures 3 to 5. In figure 3 the mobile communication device 201 provides the network communication entity 202 with the destination information and in figure 4 the network communication entity 202 provides the mobile communication device 201 with information for refining the destination information.

Figure 5 shows a schematic diagram illustrating an interaction process 500 between the mobile communication device 201 and the communication network entity 202 according to an embodiment.

The first part of the process 500 includes steps to update the communication network entity 202 in the cellular communication network 200 with information from the mobile communication device 201 , when the mobile communication device 201 moves into the network cell(s) covered by the communication network entity 202, i.e. the current network cell. In a step 501 the communication network entity 202 sends the mobile communication device 201 a request for destination information and context information, wherein the destination information comprises the source S of the trip and the destination likelihood distribution p(D) of the mobile communication device 201 . In a step 503 the mobile communication device 201 provides the communication network entity 202 with the destination information, wherein the destination information comprises the source of the trip S and the destination likelihood distribution p(D) of the mobile

communication device 201 . In a step 505 the communication network entity 202 computes a Bayesian a posteriori estimate assuming the destination likelihood distribution p(D) is correct, and adds a weight p(D) to a current count∑τ p^*(T,H,L,S-D) of a past trajectory H of the mobile communication device 201 under a context comprising a current position L of the mobile communication device 201 , the source information S, and the destination likelihood distribution p(D).

Moreover, the communication network entity 202 adds a weight p(D) to the current count p^*(T,H,L,S-D), once the mobile communication device 201 has moved and the future local trajectory T has been observed.

The second part of the process 500 includes steps to update the mobile communication device 201 with the data determined by the communication network entity 202. These steps could be triggered by the mobile communication device 201 requesting feedback from the communication network entity 202. Such a request for feedback can be included in the transmission of the destination in step 503. Responsive to the request for feedback by the mobile communication device 201 the communication network entity 202 computes a refined destination likelihood distribution on the basis of following equations: p^*(D I H,L,S) = ∑T p^*(T,H,L,S-D) /∑_D,_T p^*(T,H,L,S-D), wherein∑τ denotes the sum over all recorded trajectories, p^*(T,H,L,S-D) denotes the distribution over the set of recorded trajectories T between the current position L and the likely destination D of the mobile communication device 201 , and∑D,T denotes the sum over all destinations and all recorded trajectories.

In a step 507 the communication network entity 202 sends the refined destination likelihood distribution p^*(D | H,L,S) to the mobile communication device 201 , wherein the refined destination likelihood distribution can be combined with the destination likelihood distribution p(D) by the mobile communication device 201 , or the communication network entity 202 directly sends a combination of the refined destination likelihood distribution and the destination likelihood distribution p(D) to the mobile communication device 201 .

The destination likelihood distribution p(D) can be updated by the following equation: p(D) <- p(D) p^*(D I H,L,S) /∑_D p(D) p^*(D | H,L,S), wherein∑D denotes the sum over all destination.

Figure 6 shows a schematic diagram of a method 600 of predicting mobility of a mobile communication device, such as the mobile communication device 201 shown in figure 2 in the spatial area covered by the cellular communication network 200.

The method 600 comprises the following steps: determining 601 position information of the mobile communication device 201 , wherein the position information comprises a current position L of the mobile communication device 201 in the current network cell; providing 603 destination information of the mobile communication device 201 on the basis of the current position L of the mobile communication device 201 , wherein the destination information comprises at least one likely destination D of the mobile communication device 201 and/or a destination likelihood distribution p(D) defining at least one likely destination D of the mobile communication device 201 ; and predicting 605 at least a portion of a future trajectory of the mobile communication device 201 between the current position L and the likely destination D on the basis of a subset of a set of recorded local trajectories T, wherein each trajectory of the subset of the set of recorded local trajectories is associated with the likely destination D of the mobile communication device 201 .

Figure 7 shows a schematic diagram of simulation results for the mobile communication device 201 , the communication network entity 202 and the method 600 according to an embodiment based on a randomly generated network 700. In the randomly generated network 700, nodes represent different cells of the cellular communication network 200, and edges represent possible transition trajectories between nearby cells. It is assumed that each node has its own communication network entity 202 and can predict which edge (i.e., a possible transition trajectory of the mobile communication device 201 between a source position S and a likely destination D) will be followed by the mobile communication device 201 to leave the node.

Mobility processes for one thousand mobile communication devices 201 have been generated in the randomly generated network 700, wherein each mobile communication device 201 has a set of ten likely destinations (i.e., points of interests), and moves from one to another according to some simple mobility processes (e.g., Markovian model of order one), and wherein each mobile communication device 201 may take different trajectories (i.e., different sequence of edges), for unknown reasons (e.g., different congestion on the roads), but short routes are preferred than long ones in general. Such a randomly generated network 700 is considered to be a reasonable modeling of the mobility processes of the mobile communication devices 201 in the cellular communication network for the purpose of evaluating the performance of embodiments of the invention.

In the randomly generated network 700 shown in figure 7, eight multiple trajectories taken by the mobile communication devices 201 between a specific source position S and a likely destination D are indicated by black bold lines.

Figure 8 shows a schematic diagram illustrating performances of different embodiments of the invention in comparison with existing approaches to mobility prediction in the randomly produced network 700 of figure 7. Mobility processes for one thousand mobile

communication devices 201 have been generated in the randomly generated network 700, wherein each mobile communication device 201 has a set of ten likely destinations, (i.e., points of interests), and moves from one to another according to some simple mobility process (e.g., Markovian model of order one), and wherein each mobile communication device 201 may take different trajectories (i.e., different sequence of edges), for unknown reasons (e.g., different congestion on the roads), but short routes are preferred than long ones in general.

As can be taken from the detailed view in figure 8, the dashed lines indicate existing approaches to mobility prediction and the solid lines 801 a and 801 b indicate the performance of embodiments of the invention. More specifically, the fraction of correct predictions provided by existing approaches, where the mobility process is not decomposed and coordinated between the mobile communication devices 201 and the communication network entity 202, is lower than that provided by embodiments of the invention, because the existing approaches fail to handle too many different possible trajectories the mobile communication device 201 could follow.

An appropriate approach to a mobility prediction should learn when it should aggregate data, such as the source information S, the destination likelihood distribution p(D), or any other relevant context, from the multiple mobile communication devices 201 to leverage statistical benefits to the mobility prediction and when it should not. The existing approaches to mobility prediction would typically either not aggregate the data at all from multiple mobile

communication devices 201 (or not enough), and thus they would need very long training periods before they learn all the possible trajectories that a mobile communication device 201 may follow, or existing approaches to mobility prediction from the previous studies would aggregate too much data, which harms the accuracy of the predictions.

As can be taken from figure 8, the fraction of correct predictions rises as more trajectories are observed, since the mobile communication device 201 and the communication network entity 202 can aggregate data of the mobility processes, improving the prediction accuracy of the future mobility of the mobile communication device 201.

While a particular feature or aspect of the disclosure may have been disclosed with respect to only one of several implementations or embodiments, such feature or aspect may be combined with one or more other features or aspects of the other implementations or embodiments as may be desired and advantageous for any given or particular application. Furthermore, to the extent that the terms "include", "have", "with", or other variants thereof are used in either the detailed description or the claims, such terms are intended to be inclusive in a manner similar to the term "comprise". Also, the terms "exemplary", "for example" and "e.g." are merely meant as an example, rather than the best or optimal. The terms "coupled" and "connected", along with derivatives may have been used. It should be understood that these terms may have been used to indicate that two elements cooperate or interact with each other regardless whether they are in direct physical or electrical contact, or they are not in direct contact with each other.

Although specific aspects have been illustrated and described herein, it will be appreciated by those of ordinary skill in the art that a variety of alternate and/or equivalent

implementations may be substituted for the specific aspects shown and described without departing from the scope of the present disclosure. This application is intended to cover any adaptations or variations of the specific aspects discussed herein. Although the elements in the following claims are recited in a particular sequence with corresponding labeling, unless the claim recitations otherwise imply a particular sequence for implementing some or all of those elements, those elements are not necessarily intended to be limited to being implemented in that particular sequence.

Many alternatives, modifications, and variations will be apparent to those skilled in the art in light of the above teachings. Of course, those skilled in the art readily recognize that there are numerous applications of the invention beyond those described herein. While the present invention has been described with reference to one or more particular embodiments, those skilled in the art recognize that many changes may be made thereto without departing from the scope of the present invention. It is therefore to be understood that within the scope of the appended claims and their equivalents, the invention may be practiced otherwise than as specifically described herein.

Claims

1 . A method (600) of predicting mobility of a mobile communication device (201 ) in a cellular communication network (200), the cellular communication network (200) comprising a plurality of network cells, including a current network cell, the method (600) comprising: determining (601 ) position information of the mobile communication device (201 ), wherein the position information comprises a current position L of the mobile communication device (201 ) in the current network cell; providing (603) destination information of the mobile communication device (201 ) on the basis of the current position L of the mobile communication device (201 ), wherein the destination information comprises at least one likely destination D of the mobile

communication device (201 ) and/or a destination likelihood distribution p(D) defining at least one likely destination D of the mobile communication device (201 ); and predicting (605) at least a portion of a future trajectory of the mobile communication device (201 ) between the current position L and the likely destination D on the basis of a subset of a set of recorded local trajectories T, wherein each trajectory of the subset of the set of recorded local trajectories is associated with the likely destination D of the mobile communication device (201 ).

2. The method (600) of claim 1 , wherein each trajectory of the subset of the set of recorded local trajectories comprises a first portion, wherein the first portion coincides at least within the current network cell with a global trajectory between the current position L and the likely destination D of the mobile communication device (201 ).

3. The method (600) of claim 1 or 2, wherein the position information further comprises a previous position S of the mobile communication device (201 ) and each trajectory of the subset of the set of recorded local trajectories further comprises a second portion between the previous position S and the current position L.

4. The method (600) of any one of the preceding claims, wherein the position information further comprises a past trajectory H of the mobile communication device (201 ) between a previous position S and the current position L of the mobile communication device (201 ) and each trajectory of the subset of the set of recorded local trajectories further comprises a second portion corresponding to the past trajectory H of the mobile communication device (201 ) between the previous position S and the current position L of the mobile communication device (201 ).

5. The method (600) of claim 4, wherein the step of predicting (605) at least a portion of the future trajectory of the mobile communication device (201 ) between the current position L and the likely destination D on the basis of the subset of the set of recorded local trajectories T comprises the step of determining the following conditional likelihood distribution: p^*(T I H.L.S) =∑D p(D) p^*(T I H,L,S-D), wherein∑D denotes the sum over all destinations and wherein the conditional distribution p^*(T I H,L,S-D) is based on the following equation: p^*(T I H,L,S-D) = p^*(T,H,L,S-D) /∑_T p^*(T,H,L,S-D), wherein p^*(T,H,L,S-D) denotes the distribution over the set of recorded local trajectories T and∑T denotes the sum over the set of recorded local trajectories T.

6. The method (600) of claim 4 or 5, wherein the past trajectory H of the mobile communication device (201 ) is the past trajectory of the mobile communication device (201 ) in the current network cell of the cellular communication network (200).

7. The method (600) of any one of claims 4 to 6, wherein the method (600) comprises the further step of refining the destination information of the mobile communication device (201 ) on the basis of the past trajectory H between the previous position S and the current position L of the mobile communication device (201 ).

8. The method (600) of claim 7, wherein the step of refining the destination information comprises the step of determining a refined destination likelihood distribution on the basis of the following equations: p^*(D I H.L.S) =∑T p^*(T,H,L,S-D) /∑_D,_T p^*(T,H,L,S-D), p(D) <- p(D) p^*(D I H.L.S) /∑_D p(D) p^*(D | H,L,S) wherein∑D denotes the sum over all destinations,∑τ denotes the sum over all recorded local trajectories and∑D,T denotes the sum over all destinations and all recorded local trajectories.

9. The method (600) of any one of the preceding claims, wherein the future trajectory of the mobile communication device (201 ) between the current position L and the most likely destination D is predicted on the basis of the subset of the set of recorded local trajectories T associated with the likely destination D by selecting the trajectory of the subset of the set of recorded local trajectories T that occurs most often in the subset of the set of recorded local trajectories T.

10. The method (600) of any one of the preceding claims, wherein the step of providing (603) destination information comprises the step of selecting the likely destination D of the mobile communication device (201 ) from a set of points of interest associated with the mobile communication device (201 ) on the basis of the current position L of the mobile

communication device (201 ) and/or the step of determining the destination likelihood distribution p(D) using a set of points of interest associated with the mobile communication device (201 ) on the basis of the current position L of the mobile communication device (201 ).

1 1 . The method (600) of any one of the preceding claims, wherein the method (600) comprises the further step of providing context information associated with the mobile communication device (201 ) and wherein the step of predicting (605) at least a portion of the future trajectory of the mobile communication device (201 ) comprises predicting at least a portion of the future trajectory of the mobile communication device (201 ) between the current position L and the likely destination D on the basis of the subset of the set of recorded local trajectories T and on the basis of the context information.

12. A mobile communication device (201 ) for communication in a cellular communication network (200), the cellular communication network (200) comprising a plurality of network cells, including a current network cell, wherein the mobile communication device (201 ) comprises: a communication interface (201 a) for communication with a communication network entity (202) of the cellular communication network (200); and a processor (201 b) configured to: provide destination information of the mobile communication device (201 ) on the basis of a current position L of the mobile communication device (201 ) in the current network cell and/or a previous position S of the mobile communication device (201 ), wherein the destination information comprises at least one likely destination D of the mobile communication device (201 ) and/or a destination likelihood distribution p(D) defining at least one likely destination D of the mobile communication device (201 ); provide the destination information via the communication interface (201 a) to the

communication network entity (202); and refine the destination information of the mobile communication device (201 ) on the basis of refined destination information provided by the communication network entity (202) via the communication interface (201 a), wherein the refined destination information comprises at least a portion of a future trajectory of the mobile communication device (201 ) between the current position L and the likely destination D, a refined at least one likely destination D of the mobile communication device (201 ) and/or a refined destination likelihood distribution p(D) defining at least one likely destination D of the mobile communication device (201 ).

13. A communication network entity (202) for predicting mobility of a mobile

communication device (201 ) in a cellular communication network (200), the cellular communication network (200) comprising a plurality of network cells, including a current network cell, wherein the communication network entity (202) comprises: a communication interface (202a) for receiving destination information from the mobile communication device (201 ), wherein the destination information comprises at least one likely destination D of the mobile communication device (201 ) and/or a destination likelihood distribution p(D) defining at least one likely destination D of the mobile communication device (201 ); a memory (202c) configured to store a set of recorded local trajectories T; a processor (202b) configured to determine position information of the mobile communication device (201 ), wherein the position information comprises a current position L of the mobile communication device (201 ) in the current network cell, and to predict at least a portion of a future trajectory of the mobile communication device (201 ) between the current position L and the likely destination D on the basis of a subset of the set of recorded local trajectories T, wherein each trajectory of the subset of the set of recorded local trajectories is associated with the likely destination of the mobile communication device (201 ).

14. The communication network entity (202) of claim 13, wherein each trajectory of the subset of the set of recorded local trajectories comprises a first portion, wherein the first portion coincides at least within the current network cell with a global trajectory between the current position L and the likely destination D of the mobile communication device (201 ).

15. The communication network entity (202) of claims 13 or 14, wherein the position information further comprises a previous position S of the mobile communication device (201 ) and each trajectory of the subset of the set of recorded local trajectories further comprises a second portion between the previous position S and the current position L of the mobile communication device (201 ).

16. The communication network entity (202) of any one of claims 13 to 15, wherein the position information further comprises a past trajectory H of the mobile communication device (201 ) between a previous position S and the current position L of the mobile communication device (201 ) and each trajectory of the subset of the set of recorded local trajectories further comprises a second portion corresponding to the past trajectory H of the mobile communication device (201 ) between the previous position S and the current position L of the mobile communication device (201 ).

17. The communication network entity (202) of any one of claims 13 to 16, wherein the processor is further configured to determine a refined estimate of the likely destination D of the mobile communication device (201 ) and to feed the refined estimate of the likely destination D via the communication interface (202a) back to the mobile communication device (201 ) and/or the processor is further configured to determine a refined destination likelihood distribution p(D) and to feed the refined destination likelihood distribution p(D) via the communication interface (202a) back to the mobile communication device (201 ).

18. The communication network entity (202) of any one of claims 13 to 17, wherein the memory (202c) is configured to store each local trajectory of the set of recorded local trajectories T together with the most likely destination D associated therewith.

19. A computer program comprising program code for performing the method of any one of claims 1 to 1 1 when executed on a computer.