KR20230157837A - Operating method of communicaiton devices for paging and communicaiton devices thereof - Google Patents

Abstract

A method of operating a data analysis device including first data analysis devices distributed in a network and a centralized second data analysis device according to an embodiment includes the first data analysis devices receiving a user terminal from mobility management devices in the network. Mobility data is collected, a second data analysis device preprocesses the mobility data, and the mobility data preprocessed by the second data analysis device is applied to a neural network-based prediction model, so that the prediction model is a target location where the user terminal is expected to move. The base stations serving can be trained to predict, and the trained prediction model and identification information of the user terminal can be delivered to the first data analysis devices.

Description

Method of operating communication devices for paging and the communication devices {OPERATING METHOD OF COMMUNICAITON DEVICES FOR PAGING AND COMMUNICAITON DEVICES THEREOF}

The disclosure below relates to communication devices and methods of operating communication devices for paging.

In a mobile network, a base station can repeat the process of disconnecting and reconnecting a wireless connection to reduce power consumption of a user equipment (UE) and save wireless resources. For example, a base station in a 5G network, such as a next generation node B (gNB), may disconnect the connection between the base station and the user terminal if there is no traffic between the user terminal and the base station for a certain period of time. As the connection with the base station is released, the operation mode of the user terminal may change from active mode to idle mode. When downlink data to be transmitted to a user terminal in idle mode is generated, an access and mobility management function (AMF) device may transmit a paging message to multiple base stations to find the location of the user terminal.

Paging can cause a lot of signaling traffic in the core network. For example, as the size of a cell becomes smaller, such as in a 5G network, more signaling for paging occurs, so paging may cost more. Additionally, delay due to paging makes it difficult to satisfy services with high latency requirements, such as ultra-reliable low latency communications (URLLC).

Various embodiments consider that the location of the user terminal varies depending on the time elapsed in the idle mode, and include not only location information that the user terminal has moved in the active mode, but also information on the time elapsed in the user terminal in the idle mode. By applying it, you can train a prediction model that predicts the mobility of a user terminal suitable for paging.

Various embodiments may predict the location of a user terminal for paging by applying mobility data of the user terminal to a trained prediction model, and perform paging by a base station serving the predicted location.

In various embodiments, the ratio of base stations to be used for multi-step paging can be adjusted by determining the paging service type corresponding to the user terminal according to the service type and/or charging policy of the user terminal.

According to one embodiment, a method of operating a data analysis device including first data analysis devices distributed in a network and a second data analysis device centralized includes: the first data analysis devices receiving information from mobility management devices in the network; An operation of collecting mobility data of a user terminal, an operation of preprocessing the mobility data by the second data analysis device, the second data analysis device applying the preprocessed mobility data to a neural network-based prediction model, Training a prediction model to predict base stations serving a target location where the user terminal is expected to move, and transmitting the trained prediction model and identification information of the user terminal to the first data analysis devices. may include.

According to one embodiment, a method of operating a mobility management device includes transmitting mobility data of a user terminal to a first data analysis device distributed in a network, where the first data analysis device includes a trained prediction model based on a neural network. An operation of receiving a base station list including base stations serving a target location where the user terminal is expected to move, predicted by the first data analysis device by applying the mobility data to the prediction model, in the base station list It may include an operation of determining a target base station that performs paging for each stage of multi-stage paging among the included base stations, and an operation of performing the multi-stage paging for the user terminal by the target base station.

According to one embodiment, a data analysis device including first data analysis devices distributed in a network and a second data analysis device centralized includes a communication interface for collecting mobility data of a user terminal from mobility management devices in the network; And a processor that pre-processes the mobility data, applies the pre-processed mobility data to a neural network-based prediction model, and trains the prediction model to predict base stations serving a target location where the user terminal is expected to move. , the communication interface may transmit the base station list predicted by the trained prediction model and the identification information of the user terminal to the mobility management devices.

According to one embodiment, the mobility management device transmits mobility data of the user terminal to a first data analysis device distributed in the network, where the first data analysis device includes a trained prediction model based on a neural network, and the mobility A communication interface for receiving a base station list including base stations serving a target location to which the user terminal is expected to move, predicted by the prediction model by the first data analysis device based on data, and a multi-stage of the base stations It may include a processor that determines a target base station that performs paging for each stage of paging, and performs the multi-stage paging for the user terminal by the target base station.

A communication device according to an embodiment can predict the mobility of a user terminal suitable for paging with higher accuracy by training a neural network-based prediction model by reflecting the elapsed time in the idle mode of the user terminal.

A communication device according to an embodiment may reduce signaling traffic and/or delay time due to paging by performing multi-stage paging by base stations predicted using a trained prediction model.

The communication device according to one embodiment determines the paging service type corresponding to the user terminal according to the service type and/or charging policy of the user terminal, and determines the proportion of base stations that perform paging for each stage of multi-stage paging according to the paging service type. By adjusting it, a paging service tailored to the service requested by the user terminal can be provided.

In addition, various effects that can be directly or indirectly identified through this document may be provided.

Figure 1 is a diagram showing the structure of a 5G mobile communication system according to an embodiment.
FIG. 2 is a diagram for explaining a tracking area (TA) and a tracking area identifier (TAI) in a 5G mobile communication system according to an embodiment.
Figure 3 is a diagram for explaining a typical multi-step paging method.
FIG. 4 is a diagram conceptually illustrating a paging method applying mobility prediction of a user terminal according to an embodiment.
Figure 5 is a flowchart showing a method of operating a data analysis device according to an embodiment.
FIG. 6 is a diagram illustrating a process of preprocessing mobility data according to an embodiment.
Figure 7 is a diagram for explaining the structure and operation of a neural network-based prediction model according to an embodiment.
Figure 8 is a flowchart showing a method of operating a mobility management device according to an embodiment.
Figure 9 is a diagram for explaining multi-step paging methods according to an embodiment.
FIG. 10 is a diagram illustrating operations in an offline training process and an online prediction process between a data analysis device and a mobility management device according to an embodiment.
Figure 11 is a block diagram of a data analysis device according to one embodiment.
Figure 12 is a block diagram of a mobility management device according to one embodiment.
FIG. 13 is a block diagram of an electronic device that is an example of a user device in a network environment according to various embodiments.

Hereinafter, embodiments will be described in detail with reference to the attached drawings. In the description with reference to the accompanying drawings, identical components will be assigned the same reference numerals regardless of the reference numerals, and overlapping descriptions thereof will be omitted.

Electronic devices according to various embodiments disclosed in this document may be of various types. Electronic devices may include, for example, portable communication devices (e.g., smartphones), computer devices, portable multimedia devices, portable medical devices, cameras, wearable devices, or home appliances. Electronic devices according to embodiments of this document are not limited to the above-described devices.

The various embodiments of this document and the terms used herein are not intended to limit the technical features described in this document to specific embodiments, and should be understood to include various changes, equivalents, or replacements of the embodiments. In connection with the description of the drawings, similar reference numbers may be used for similar or related components. The singular form of a noun corresponding to an item may include one or more of the above items, unless the relevant context clearly indicates otherwise. As used herein, “A or B”, “at least one of A and B”, “at least one of A or B”, “A, B or C”, “at least one of A, B and C”, and “A Each of phrases such as “at least one of , B, or C” may include any one of the items listed together in the corresponding phrase, or any possible combination thereof. Terms such as "first", "second", or "first" or "second" may be used simply to distinguish one element from another, and may be used to distinguish such elements in other respects, such as importance or order) is not limited. One (e.g. first) component is said to be "coupled" or "connected" to another (e.g. second) component, with or without the terms "functionally" or "communicatively". Where mentioned, it means that any of the components can be connected to the other components directly (e.g. wired), wirelessly, or through a third component.

The term “module” used in various embodiments of this document may include a unit implemented in hardware, software, or firmware, and is interchangeable with terms such as logic, logic block, component, or circuit, for example. It can be used as A module may be an integrated part or a minimum unit of the parts or a part thereof that performs one or more functions. For example, according to one embodiment, the module may be implemented in the form of an application-specific integrated circuit (ASIC).

Various embodiments of this document are stored in a storage medium (e.g., internal memory 136 or external memory 138) that can be read by a machine (e.g., electronic device 101 of FIG. 1). It may be implemented as software (e.g., program 140) including one or more instructions. For example, a processor (e.g., processor 120) of a device (e.g., electronic device 101) may call at least one command among one or more commands stored from a storage medium and execute it. This allows the device to be operated to perform at least one function according to the at least one instruction called. The one or more instructions may include code generated by a compiler or code that can be executed by an interpreter. A storage medium that can be read by a device may be provided in the form of a non-transitory storage medium. Here, 'non-transitory' only means that the storage medium is a tangible device and does not contain signals (e.g. electromagnetic waves), and this term refers to cases where data is semi-permanently stored in the storage medium. There is no distinction between temporary storage cases.

According to one embodiment, methods according to various embodiments disclosed in this document may be included and provided in a computer program product. Computer program products are commodities and can be traded between sellers and buyers. The computer program product is distributed in the form of a machine-readable storage medium (e.g. compact disc read only memory (CD-ROM)), or through an application store (e.g. Play StoreTM) or on two user terminals (e.g. It can be distributed (e.g. downloaded or uploaded) directly between smart phones) or online. In the case of online distribution, at least a portion of the computer program product may be at least temporarily stored or temporarily created in a machine-readable storage medium, such as the memory of a manufacturer's server, an application store's server, or a relay server.

According to various embodiments, each component (e.g., module or program) of the above-described components may include a single or plural entity, and some of the plurality of entities may be separately placed in other components. there is. According to various embodiments, one or more of the components or operations described above may be omitted, or one or more other components or operations may be added. Alternatively or additionally, multiple components (eg, modules or programs) may be integrated into a single component. In this case, the integrated component may perform one or more functions of each component of the plurality of components in the same or similar manner as those performed by the corresponding component of the plurality of components prior to the integration. . According to various embodiments, operations performed by a module, program, or other component may be executed sequentially, in parallel, iteratively, or heuristically, or one or more of the operations may be executed in a different order, or omitted. Alternatively, one or more other operations may be added.

Figure 1 is a diagram showing the structure of a 5G mobile communication system according to an embodiment. Referring to FIG. 1, the 5G mobile communication system 100 includes, for example, a user equipment (UE) 110 (e.g., the electronic device 1301 of FIG. 13) and a 5G radio access network (RAN) ( 120), AMF (access and mobility management function) (130), AUSF (authentication server function) (140), UDM (user data management) (150), SMF (session management function) (160), PCF (policy control function) ) (170), UPF (user plane function) 180, and DN (data network) 190, but is not necessarily limited thereto. Hereinafter, the 5G RAN 120 may include a 'base station'.

For the components of the 5G mobile communication system 100, each component and the connection interface between them may be defined in units of network functions (NF). As a result, the functions of each component can be implemented through virtualization using physical network equipment or network resources. First, when a 5G subscriber first accesses the 5G network through the user terminal 110, the AMF 130 can authenticate the user terminal 110 using the subscriber's authentication data stored in the AUSF 140. Here, in addition to subscriber authentication, the AMF 130 may be responsible for mobility management that supports uninterrupted communication even if the user terminal 110 moves to the cell area of another base station. At least some of the other subscription information, rate plans, and charging policies can be managed separately through a database called UDM (150).

After subscriber authentication, the user terminal 110 establishes a connection with the 5G RAN 120, which is the base station in the current cell area, through a wireless channel, and finally connects to the DN 190 corresponding to the Internet network through the UPF 180. can be connected Here, the DN 190 includes the Internet, and the UPF 180 can serve as a gateway for packets transmitted through the 5G RAN to be finally delivered to the DN 190.

When the user terminal 110 connected to the DN 190 receives multiple multi-services at the same time, it may have multiple connection sessions. In this case, an SMF 160 is allocated to each session and can be managed to prevent service interruption due to disconnection of each session. PCF 170 may determine and apply policies for session management through SMF 160 and mobility management through AMF 130. Policies for mobility and session management are determined through the PCF (170), and each policy can be transferred from the PCF (170) to the AMF (130) and SMF (160) and applied.

The user terminal 110, 5G RAN 120, UPF 180, and DN 190, which are the sections through which actual data flows, are called the 'user plane' 101, and the 5G mobile communication system 100 AMF (130), AUSF (140), UDM (150), SMF (160), and PCF (170), which are areas for control and management, may be called a 'control plane' (103).

FIG. 2 is a diagram for explaining a tracking area (TA) and a tracking area identifier (TAI) in a 5G mobile communication system according to an embodiment. Referring to FIG. 2, user terminals 250 and 260 (e.g., user terminal 110 of FIG. 1 and/or electronic device 1301 of FIG. 13) according to an embodiment are served by base stations 215, 225, 235, and 245. A diagram 200 is shown showing the process of moving each cell 210, 220, 230, and 240.

For example, when the user terminals 250 and 260 are in an active state performing communication, the network can know the locations of the user terminals 250 and 260 on a cell basis. However, when the user terminals 250 and 260 are in an idle state not performing communication, the network can know the locations of the user terminals 250 and 260 on a TA (tracking area) basis rather than a cell basis. A single TA can be defined as a group of multiple adjacent base stations, and each of the base stations 215, 225, 235, and 245 can know which TA it belongs to.

For example, if downlink traffic toward the user terminals 250 and 260 occurs while the user terminals 250 and 260 are in an idle state, the network may transmit a paging message to the user terminals 250 and 260. . At this time, the network transmits a paging message to the base stations of the TA to which the corresponding user terminals 250 and 260 belong, and each of the base stations transmits the paging message to its own coverage, that is, to the area served by each base station. User terminals 250 and 260 that are in an idle state can be woken up.

The tracking area identifier (TAI) 270 is a unique identifier that identifies the TA and may include, for example, a public land mobile network (PLMN) ID and a tracking area identifier (TAI). The PLMN ID is a unique network identifier assigned to a mobile communication service provider and may include a mobile country code (MCC) and a mobile network code (MNC). TAI 270 may have a unique value assigned to each TA by the mobile communication service provider.

For example, the first cell 210 may correspond to TA1, the second cell 220 may correspond to TA2, and the third cell 230 and fourth cell 240 may correspond to TA3. The first cell 210 may be served by a first base station (gNB1) 215, and the second cell 220 may be served by a second base station (gNB2) 225. Additionally, the third cell 230 may be served by a third base station (gNB3) 235, and the fourth cell 240 may be served by a fourth base station (gNB4) 215.

When the TA changes as the first user terminal (UE 1) 250 moves through the first cell 210, the second cell 220, and the third cell 230, the first user terminal 250 Can report a registration request message notifying the fact that the TA has changed to the network. In response to the registration request message, the first user terminal 250 may receive a TAI list - for example, {TAI1, TAI2} - from the network. At this time, the first user terminal 250 does not send a registration request message when moving between TA1 and TA2 included in the TA list, but sends a registration request message when moving to a location other than TA1 and TA2 (for example, TA3). You can send a message. When the first user terminal 250 moves to TA3, the first user terminal 250 receives a TAI list (e.g., {TAI2, TAI3}) from the network and updates the TAI list according to the location movement. You can.

Figure 3 is a diagram for explaining a typical multi-step paging method. 3, a diagram 300 is shown illustrating multi-stage paging performed by a mobility management device 310 in a network.

Multi-level paging performed in commercial LTE and/or commercial 5G core may be performed by a mobility management device. Mobility management device 310 may include, but is not necessarily limited to, an access and mobility management function (AMF) device and/or a mobility management entity (MME), for example.

As shown in FIG. 3, the multi-stage paging method spans three stages such as first stage paging (1 ^st paging), second stage paging (2 ^nd paging), and third stage paging (3 ^rd paging). It can proceed. The mobility management device 310 selects the last confirmed user terminal (e.g., the user terminal 110 of FIG. 1, the user terminals 250 and 260 of FIG. 2, and/or the electronic device of FIG. 13) in each of the three stages. Based on the location of (1301)), the paging area can be gradually expanded to the base station, TA (Tracking Area), and TAL (Tracking Area List).

In the first stage of paging, a paging message may be transmitted to the last known base station 330 of the user terminal. If the user terminal is no longer in the last known cell of base station 330, mobility management device 310 may extend the paging area to TA 350 in a second stage of paging. Likewise, if the second stage paging also fails, the mobility management device 310 may expand the paging area from the TA 350 to the TA list (TAL) 370. Paging failure can cause an exponential increase in paging messages that are sent redundantly and a linear increase in delay for localizing the user terminal. 5G mobile networks have a larger number of base stations than fourth generation (4G) mobile networks due to smaller cell sizes, which may increase the transmission of redundant paging messages in 5G mobile networks. As such, if paging fails at each stage of multi-stage paging, a delay equal to the paging response waiting time occurs, so signaling traffic and delay time may increase as each stage of paging progresses.

FIG. 4 is a diagram conceptually illustrating a paging method applying mobility prediction of a user terminal according to an embodiment.

Compared to previous networks, the 5G mobile network not only has a smaller cell size, resulting in a greater number of paging signaling, but also can include services with high latency requirements, such as URLLC, for example. Therefore, 5G mobile networks can reduce signaling traffic by performing paging by fewer base stations compared to previous networks, and reduce latency due to paging by achieving a higher paging success rate.

Referring to FIG. 4, the activation mode of a user terminal (e.g., the user terminal 110 of FIG. 1, the user terminals 250 and 260 of FIG. 2, and/or the electronic device 1301 of FIG. 13) according to an embodiment. A diagram 400 is shown showing mobility in 410 and mobility in idle mode 430.

The mobility of the user terminal for paging is determined by the user terminal's movement path (e.g., location 1 (411) -> location 2 (413) -> location 3 (415) as determined in the activation mode (410) in which the user terminal is performing communication. -> .. -> It can be determined by predicting the location of the user terminal at the time of paging based on location n (417)). However, in the idle mode 430 in which the user terminal does not perform communication, for example, location n+1 (431) -> location n+2 (433) -> location n+3 (435) -> .. Likewise, when the user moves, the location n+m (450) of the user terminal may vary depending on the elapsed time of the idle mode (430). Therefore, if the user's movement path is predicted without considering the elapsed time in the idle mode 430, the accuracy of the predicted result may be low and may not be sufficient to be applied to paging.

In one embodiment, the prediction accuracy of the prediction model is improved by predicting the user's mobility by learning the elapsed time in the idle mode 430 in addition to the movement path of the user terminal in the active mode 410, while improving the prediction accuracy of the prediction model. and/or delay time may be reduced. In one embodiment, for example, by predicting the mobility of a user terminal suitable for paging by reflecting the elapsed time in the idle mode 430 in the machine learning model, base stations serving the user terminal at the time of paging can also be predicted.

Additionally, in one embodiment, multi-stage paging is performed depending on whether the paging service type corresponding to the user terminal is a first service type (iPRS) for reducing signaling or a second service type (iPRD) for reducing delay. Signaling traffic and delay time can be provided according to the service type by adjusting the proportion of predicted base stations to be used in each step. The paging service type according to one embodiment will be described later with reference to Figure 9 below.

Figure 5 is a flowchart showing a method of operating a data analysis device according to an embodiment. In the following embodiments, each operation may be performed sequentially, but is not necessarily performed sequentially. For example, the order of each operation may be changed, and at least two operations may be performed in parallel.

A data analysis device (eg, the data analysis device 1100 in FIG. 11) may collect network data at the center of the entire network system and analyze the network data. A data analysis device may analyze network data, for example, for one or more network slices. A data analysis device may collaborate with a network function device of the 5G core network for network instances, where analysis is independently performed by a network function device of the 5G core network. A 'network function device' may be a device that performs various functions of the 5G core network.

Network data used in the 5G core network can be input to a data analysis device for data analysis. Data analytics devices can utilize existing network services based on other 5G core network functions and interfaces to communicate with operation administration maintenance (OAM) devices. An OAM device may be an information provider for a data analysis device, or may be a potential consumer.

Interactions between data analysis devices and network function devices may be performed by, for example, a local public land mobile network (PLMN). The reporting network function device and the data analysis device may belong to the same PLMN.

A data analysis device may access network data from a data repository, for example, a unified data repository (UDR). For network function devices in the 5G core network, the data analysis device can acquire (collect) network data, analyze the collected network data, and provide analysis results of the network data to the network function devices and/or OAM device. there is.

The data analysis device may be, for example, a network data analytics function (NWDAF), but is not necessarily limited thereto. Data analysis devices can be located on any device within the 5G core network.

Referring to the flowchart 500 of FIG. 5, the data analysis device according to one embodiment includes first data analysis devices distributed within a network (e.g., first data analysis devices 1003 and 1060 of FIG. 10) and a central It may include a second focused data analysis device (e.g., the second data analysis device 1001 of FIG. 10) and may perform operations 510 to 540.

In operation 510, the first data analysis devices receive a user terminal (e.g., user terminal 110 of FIG. 1, user terminals 250 and 260 of FIG. 2, and/or electronic device of FIG. 13) from mobility management devices in the network. (1301)) mobility data can be collected.

Hereinafter, for convenience of explanation, 'mobility data of user terminal' may be simplified and expressed as 'mobility data'. The mobility data may be, for example, mobility data of the user terminal collected by mobility management devices distributed within the network, such as an AFM device, transmitted to the first data analysis devices, but is not necessarily limited thereto.

Mobility data includes, for example, location information that the user terminal has moved in an active mode in which communication is performed, a first elapsed time in which the user terminal has elapsed in an active mode, and a first elapsed time in an idle mode in which the user terminal has not performed communication. 2 It may include any one of the elapsed times or a combination thereof, but is not necessarily limited thereto. The location information includes identification information of the user terminal (e.g., ID of the user terminal), information about the first base station corresponding to the first location (e.g., source) from which the user terminal departs in the activation mode, and the user terminal. In the activation mode, it may include any one or a combination of information about the second base station corresponding to the second location (eg, destination) arrived at by moving from the first location, but is not necessarily limited thereto.

Hereinafter, in this specification, 'first elapsed time' may refer to the time elapsed when the user terminal is in an active mode, and 'second elapsed time' may refer to the time elapsed when the user terminal is in an idle mode.

Mobility management devices (e.g., AMF 130 in FIG. 1, mobility management device 310 in FIG. 3, mobility management device 805 in FIG. 8, mobility management device 1005, 1065 in FIG. 10, and/or The mobility management device 1200 of FIG. 12 may include, for example, an access and mobility management function (AMF) device and/or a mobility management entity (MME), but is not necessarily limited thereto. Mobility management devices may sample mobility data of the user terminal at fixed intervals, such as 3 seconds, 5 seconds, or 10 seconds, and transmit the sampled mobility data to the first data analysis devices. The first data analysis devices may periodically transmit the mobility data collected from the mobility management device to the second data analysis device, or may transmit it at the request of the second data analysis device.

In operation 520, the second data analysis device may preprocess the mobility data collected in operation 510. Here, 'pre-processing' refers to a form for using mobility data for learning and evaluation of a prediction model (e.g., the prediction model 700 in FIG. 7 and/or the prediction model 1050 in FIG. 10). For example, it may correspond to a process of processing in batch form that can be processed in batches. For example, the second data analysis device may divide the mobility data by identification information of the user terminal and generate one continuous sequence. The second data analysis device may divide one continuous sequence into a plurality of unit sequences including location information corresponding to a fixed time unit. The second data analysis device includes a plurality of devices based on the movement path of the user terminal during the first elapsed time during which the user terminal is in the active mode and the target location of the user terminal during the second elapsed time when the user terminal is in the idle mode. The unit sequences can be configured as a batch for learning and evaluating the prediction model.

An example of a method by which the second data analysis device configures a plurality of unit sequences in batches is as follows.

The second data analysis device may set the first unit sequences belonging to the first elapsed time among the plurality of unit sequences as input for training the prediction model. The second data analysis device may set second unit sequences belonging to the second elapsed time among the plurality of unit sequences as labels for evaluating the prediction model. The second data analysis device may include information including, for example, identification information of the user terminal, a second elapsed time, a movement path of the user terminal during the first elapsed time, and a target location of the user terminal after the second elapsed time. It can be configured in batches, but is not necessarily limited to this. The method by which the second data analysis device preprocesses mobility data will be described in more detail with reference to FIG. 6 below.

In operation 530, the second data analysis device applies the mobility data preprocessed in operation 520 to a neural network-based prediction model, so that the prediction model can be trained to predict base stations serving the target location where the user terminal is expected to move. there is. The second data analysis device may generate a prediction model to determine the target location using mobility data sampled by the mobility management device, for example, over a fixed time interval and a first elapsed time during which the user terminal is in an active mode. It can be trained to predict base stations serving.

The prediction model may be, for example, a spatiotemporal prediction model that predicts base stations serving the target location where the user terminal is expected to move at the time of paging when the user terminal switches from idle mode to active mode, and must be It is not limited to this.

The prediction model may include, for example, a stacked deep neural network (e.g., stacked recurrent neural network 710 in FIG. 7) and fully connected layers (e.g., fully connected layers 730). and is not necessarily limited to this.

The stacked deep neural network can learn spatiotemporal characteristics of the mobility of the user terminal by receiving preprocessed mobility data. A stacked deep neural network may include, for example, an input layer composed of recurrent neural network cells, and hidden layers. As the previous location paths according to the fixed time interval of the user terminal and the second elapsed time elapsed when the user terminal is in idle mode are applied to the input layer, the hidden layers are applied to the user terminal based on the previous location paths to determine the second elapsed time. The target location expected to move in elapsed time can be trained.

Fully connected layers include, for example, a base station list (e.g., base station list 750 in FIG. 7) that includes the probability of a user terminal being located for each base station from the output of a layered deep neural network, using a softmax activation function. , and/or the base station list 1006 of FIG. 10) can be output. In one embodiment, for convenience of explanation, the soft max activation function is described as an example, but the present invention is not limited to this, and various other activation functions may be used.

The data analysis device (e.g., first data analysis devices or second data analysis device) sorts the probability of the user terminal being located for each base station output through fully connected layers, and corresponds to a certain percentage of the sorted probabilities. You can output a base station list including the base stations. The structure and operation of the prediction model will be described in more detail with reference to FIG. 7 below.

In operation 540, the second data analysis device may transmit the prediction model trained in operation 530 and the identification information of the user terminal to the first data analysis devices.

The process of performing offline training and online prediction in a network including first data analysis devices, second data analysis devices, and mobility management devices according to an embodiment is described in more detail with reference to FIG. 10 below. Explain.

FIG. 6 is a diagram illustrating a process of preprocessing mobility data according to an embodiment. Referring to FIG. 6, mobility data of a user terminal (e.g., the user terminal 110 in FIG. 1, the user terminals 250 and 260 in FIG. 2, and/or the electronic device 1301 in FIG. 13) according to an embodiment. A diagram 600 is shown showing a plurality of unit sequences 610 that divide a continuous sequence based on a fixed time unit and an arrangement 630 composed of the plurality of unit sequences 610.

As described above, a mobility management device according to an embodiment (e.g., AMF 130 in FIG. 1, mobility management device 310 in FIG. 3, mobility management device 805 in FIG. 8, mobility management device in FIG. 10 (1005, 1065), and/or the mobility management device 1200 of FIG. 12) records mobility data including location information and/or time information that the user terminal moved in the activation mode to predict the mobility of the user terminal. It can be collected. Mobility data can also be called 'log' in that it includes location information and time information that occurs in the network. Mobility data initially collected by the mobility management device may include, for example, identification information of the user terminal (UE ID), source base station (source gNB), and destination base station (destination gNB), and travel time (time). You can.

The second data analysis device (e.g., the second data analysis device 1001 in FIG. 10) may divide the collected mobility data by identification information (UE ID) of the user terminal and generate one continuous sequence. The second data analysis device may divide one continuous sequence into a plurality of unit sequences 610 including location information corresponding to a fixed time unit (eg, 5 minutes (min)).

For spatio-temporal prediction, the second data analysis device divides the mobility data by the user terminal's identification information (UE ID) into one continuous sequence (e.g. ) can be divided into a plurality of unit sequences 610 by dividing them into fixed time intervals. here, represents the identification information (UE ID) of the user terminal, may indicate location information of the user terminal corresponding to a fixed time unit.

The plurality of unit sequences 610 include unit sequences 611 corresponding to a first elapsed time when the user terminal is tracked in an active mode and unit sequences corresponding to a second elapsed time when the user terminal is tracked in an idle mode ( 613) may be included.

For example, if the fixed time interval is 5 minutes, the user terminal Mobility data of ( ,1,3,14:00), ( ,3,5,14:15), ( ,5,4,14:20) can be given as 4-tuples. In this case, the second data analysis device generates a 5-minute movement sequence from 4-tuples representing the mobility of the user terminal for 20 minutes. can be created.

According to an embodiment, the second data analysis device samples tuples at fixed intervals of 5 minutes to create a movement sequence for spatiotemporal prediction. and create a movement sequence A batch for learning and evaluation of a prediction model (e.g., the prediction model 700 of FIG. 7 and/or the prediction model 1050 of FIG. 10) may be configured. Here, c may be identification information of the user terminal or identification information of the cluster to which the user terminal belongs, may represent identification information (gNB ID) of the base station serving the location of the user terminal in 5-minute increments.

For example, the user terminal can move for 6 hours in active mode and 4 hours in idle mode, for a total of 10 hours. In this case, a prediction model is created using the movement path for 6 hours ('first elapsed time') tracked in the active mode of the user terminal and the movement path for 4 hours ('second elapsed time') in idle mode. You can train.

The second data analysis device may divide one sequence corresponding to a total of 10 hours into fixed time intervals (e.g., 5 minutes) into (10 hours x 60 minutes)/5 minutes = 120 unit sequences. . The second data analysis device trains a prediction model with (6 hours It can be used by setting it as an input for . In addition, the second data analysis device predicts (4 hours It can be set and used as a label for evaluation.

The second data analysis device provides information including a movement path of the user terminal during a first elapsed time when the user terminal is in an active mode and a target location of the user terminal during a second elapsed time when the user terminal is in an idle mode. As a basis, a plurality of unit sequences can be configured into a batch 630 for learning and evaluating a prediction model.

The second data analysis device independently performs learning and evaluation for each elapsed time, for example, [identification information (UE ID) of the user terminal, elapsed time in idle mode (second elapsed time) 631 ), the movement path of the user terminal during the first elapsed time (e.g. ) (633), (target) location (635) of the user terminal after the second elapsed time] can be configured as arrangement (630). The second data analysis device may use the normalized sequence of the same length for training and testing of the prediction model. For example, the second data analysis device A plurality of unit sequences 610 such as , , … , It can be converted to a layout 630 such as and used for learning and evaluation.

The second data analysis device may store the preprocessed mobility data, that is, the batch 630, in a cloud server or a mobility data set in a separate storage.

Figure 7 is a diagram for explaining the structure and operation of a neural network-based prediction model according to an embodiment. Referring to FIG. 7, a simplified configuration diagram of a prediction model 700 (eg, prediction model 1050 of FIG. 10) according to one embodiment is shown.

The prediction model 700 may be, for example, a machine learning model or a deep neural network (DNN)-based model.

The prediction model 700 may include a stacked deep neural network 710 that learns sequential inputs and fully connected layers 730 for generating output.

The stacked deep neural network 710 receives pre-processed mobility data and processes the user terminal (e.g., the user terminal 110 in FIG. 1, the user terminals 250 and 260 in FIG. 2, and/or the electronic device 1301 in FIG. 13). ) can learn the spatiotemporal characteristics of mobility. Deep neural network (RNN) cells 715 included in the stacked deep neural network 710 may refer to an RNN-based AI model. The deep neural network cells 715 refer to the recurrent layer of the deep neural network, and the output results of the cells can be called a hidden state. The deep neural network cells 715 may be, for example, but are not necessarily limited to recurrent neural network (RNN) cells.

The stacked deep neural network 710 may be, for example, a recurrent neural network (RNN)-type neural network such as long short-term memory (LSTM) or gated recurrent units (GRU), but is not necessarily limited thereto. The stacked deep neural network 710 can also be replaced with an AI model other than RNN.

The stacked deep neural network 710 may be composed of, for example, three layers of recurrent neural network (RNN) cells 715, but is not necessarily limited thereto. In Figure 7, for convenience of explanation, a structure in which three layers are stacked is shown, but the structure is not necessarily limited thereto. Prediction model 700 may be a multi-layer neural network with a depth greater than three layers.

The first layer of the stacked deep neural network 710 may be an input layer for receiving the previous location path (Sn) of the user terminal and the second elapsed time (t) in the idle mode. The previous location path (Sn) of the user terminal input to each of the RNN cells in the input layer represents previous location information at a fixed time interval, and can be sequentially input as the length of the input sequence from S ₁ to S _n . Since the second elapsed time (t) is not time series data, the second elapsed time (t) is input equally to all RNN cells so that the stacked deep neural network 710 can be learned regardless of the input order or location. .

The second and third layers of the stacked deep neural network 710 may operate as hidden layers that are not involved in input and output. At this time, the number of layers and the number of nodes used as hidden layers in the stacked deep neural network 710 may be experimentally and/or empirically set to values suitable for the application environment.

Since the output of RNN cells is passed to the next RNN cell of the same layer and the RNN cells of the next layer, the output of the last RNN cell of the highest layer can be the final output (O) of the stacked deep neural network 710.

The output layer 750 of the prediction model 700 can output the probability of a user terminal being located for each base station (gNB) using fully connected layers 730. The fully connected layers 730 calculate the probability of a user terminal being located for each base station from the output (O) of the stacked deep neural network 710 by activation functions such as, for example, a softmax activation function. A list of base stations included can be output.

The fully connected layers 730 may receive i outputs (O) of the stacked deep neural network 710 and output hidden state values equal to j, which is the number of base stations to be predicted. Here, i may be the same as or different from n, which is the number of input values of the stacked deep neural network 710. The softmax activation function outputs normalized values between 0 and 1, and the sum of the output values can always be 1. Accordingly, each of the output values of the prediction model 700 may represent the probability that the user terminal will be located at the corresponding base station. The mobility management device that receives the probability that the user terminal is located at the corresponding base station, that is, the list of base stations ('base station list') serving the target location where the user terminal is expected to move, from the first data analysis device, determines the predicted base station ( You can perform multi-step paging by selecting the paging target.

Figure 8 is a flowchart showing a method of operating a mobility management device according to an embodiment. In the following embodiments, each operation may be performed sequentially, but is not necessarily performed sequentially. For example, the order of each operation may be changed, and at least two operations may be performed in parallel.

Referring to the flowchart 800 of FIG. 8, a mobility management device according to an embodiment (e.g., AMF 130 of FIG. 1, mobility management device 310 of FIG. 3, mobility management device 805 of FIG. 8, The mobility management devices 1005 and 1065 of FIG. 10 and/or the mobility management device 1200 of FIG. 12) manage a user terminal (e.g., the user terminal 110 of FIG. 1 and the mobility management device 1200 of FIG. 2) through operations 810 to 840. Multi-step paging can be performed on user terminals 250 and 260 and/or the electronic device 1301 of FIG. 13).

In operation 810, the mobility management device may transmit mobility data of the user terminal to first data analysis devices distributed within the network (eg, first data analysis devices 1003 and 1060 of FIG. 10). The first data analysis device may include a trained prediction model based on a neural network (eg, the prediction model 700 of FIG. 7 and/or the prediction model 1050 of FIG. 10). Mobility data includes, for example, location information moved in an active mode in which the user terminal performs communication, a first elapsed time elapsed in the active mode by the user terminal, and elapsed time in an idle mode in which the user terminal does not perform the communication. It may include any one of the second elapsed times or a combination thereof, but is not necessarily limited thereto.

In operation 820, the mobility management device creates a base station list (e.g., the base station of FIG. 7) that includes base stations serving the target location to which the user terminal is expected to move, predicted by the first data analysis device by applying mobility data to the prediction model A list 750 and/or a base station list 1006 of FIG. 10 may be received.

In operation 830, the mobility management device may determine a target base station that performs paging for each stage of multi-stage paging among base stations included in the base station list received in operation 820. For example, if it is determined that the user terminal is not in the cell of the last known base station in activation mode, the mobility management device selects the target base station among the base stations included in the base station list according to the paging service type corresponding to the user terminal. can be decided.

The mobility management device may determine the target base station by adjusting the ratio of the first target base station used for first-stage paging and the second target base station used for second-stage paging among base stations included in the base station list according to the paging service type. . More specifically, the mobility management device determines which paging service type corresponding to the user terminal is the first service type for reduction of signaling (iPRS) and the second service type for reduction of delay (iPRD). You can. For example, the mobility management device may determine which paging service type corresponds to the user terminal, based on one of a service type and a charging policy corresponding to the user terminal.

The mobility management device may adjust the ratio of the first target base station used for first-stage paging and the second target base station used for second-stage paging among base stations according to the paging service type. For example, when it is determined that the determined paging service type is the first service type, the mobility management device determines the base station last known by the user terminal in the activation mode as the first target base station, and selects the base stations included in the base station list as the second target. This can be decided by the base station. Alternatively, when it is determined that the determined paging service type is the second service type, the mobility management device may determine a number of base stations corresponding to the first ratio among base stations included in the base station list as the first target base station. The mobility management device may determine the number of base stations corresponding to the second ratio of the base stations included in the base station list, excluding the first ratio, as the second target base stations. A method of performing multi-level paging by adjusting the ratio of base stations according to the paging service type will be described in more detail with reference to FIG. 9 below.

In operation 840, the mobility management device may perform multi-step paging for the user terminal by the target base station determined in operation 830. The mobility management device may perform first-stage paging by the first target base station determined in operation 830 and perform second-stage paging by the second target base station.

Figure 9 is a diagram for explaining multi-step paging methods according to an embodiment. Referring to FIG. 9, a diagram 900 showing paging areas according to a typical multi-step paging method 910 and multi-step paging methods 930 and 950 according to paging service types according to an embodiment is shown.

Paging service types according to one embodiment may include, for example, a first service type for reducing signaling (iPRS) and a second service type for reducing delay (iPRD). The multi-level paging method 930 may be according to a first service type (iPRS) for reducing signaling, and the multi-level paging method 950 may be according to a second service type (iPRD) for reducing delay. The first service type (iPRS) corresponds to an integration platform as a service (iPaaS) for reduced signaling, and the second service type (iPRD) corresponds to reduced delay. It may correspond to an integrated platform as a service (iPaaS).

Integration Platform as a Service (iPaaS) can be a cloud-based software package that creates new applications or connects existing services and applications together to coordinate data flows. An integration platform as a service can contain routines that can interact with existing services using standard protocols and data formats. An integrated platform-as-a-service can filter data and act as a transportation hub for data transfer. For example, an integration platform as a service can request data from one service, convert the data received through the request into another data format required by another service, and transmit it.

A typical multi-stage paging method 910 involves a user terminal (e.g., the user terminal 110 in FIG. 1, the user terminals 250 and 260 in FIG. 2, and/or the electronic device 1301 in FIG. 13) performing first-stage paging. If it is determined that it does not exist in the cell of the last known base station (e.g., base station 330 in FIG. 3), the paging area is divided into 64 cells in the second stage paging and 4096 cells in the third stage paging. Since paging messages for base stations (e.g., 5G RAN in FIG. 1, base stations 215, 225, 235, 245 in FIG. 2, and/or base station 330 in FIG. 3) are gradually expanded, a large number of paging signals must be transmitted. The paging failure rate can be lowered.

In one embodiment, a base station belonging to the TA of a typical multi-stage paging method 910 is used using a deep neural network-based prediction model (e.g., the prediction model 700 of FIG. 7 and/or the prediction model 1050 of FIG. 10). A number of base stations corresponding to 20% of the number of base stations are predicted as base station(s) to perform paging, and paging is performed using the predicted base station(s) for the first service type (iPRS) and the second service type (iPRD). Signaling costs can be reduced while increasing the success rate. The base station list predicted by the prediction model (e.g., the base station list 750 in FIG. 7 and/or the base station list 1006 in FIG. 10) may include base stations serving the target location where the user terminal is expected to move. there is.

For example, when the paging service type of the user terminal is the first service type (iPRS), the mobility management device performs paging using the last known base station in the first stage paging, similar to the typical multi-stage paging method 910. can do. Additionally, in the second stage paging, the mobility management device may transmit a paging message to base stations corresponding to an area corresponding to 20% of the TA used in the typical multi-stage paging method 910.

If the paging service type corresponding to the user terminal is the first service type (iPRS), the mobility management device may transmit a page message using the base station list instead of the TA in the second stage paging. When following the first service type (iPRS), in the second stage paging, a paging message is transmitted to 80% fewer base stations than the typical multi-stage paging method (910), and paging is performed using the predicted base stations. By increasing the paging success rate, signaling costs can be greatly reduced.

Until recently, paging cost was mainly considered in terms of signaling, but in URLLC services in 5G mobile networks, paging delay may be more important to user experience.

For example, if the paging service type of the user terminal is the second service type (iPRD), the mobility management device includes the base stations included in the base station list predicted by the prediction model, along with the cell of the last known base station in the previous paging. The first stage paging is performed by base stations corresponding to 8% of the serving areas, that is, base stations serving areas corresponding to 20% of the size of the TA used in the typical multi-stage paging method 910. You can. In addition, the mobility management device uses base stations corresponding to 12% of the remaining 8% of the base stations serving areas corresponding to 20% of the TA used in the typical multi-step paging method 910, excluding the 8% previously used. Two-step paging can be performed.

When the paging service type is the second service type (iPRD), the paging success rate is increased by the increased number of base stations in the first stage paging, which has the effect of reducing delay time more than the first service type (iPRS). By increasing the number of paging signaling required in step paging, the total number of paging signaling can be increased compared to the first service type (iPRS).

Since the number of signaling and delay time in paging may be a tradeoff, the mobility management device can use the paging service type appropriate for the user depending on the service or charging policy.

The mobility device may determine which paging service type corresponds to the user terminal, for example, based on one of the service type and charging policy corresponding to the user terminal.

Depending on the embodiment, the mobility device may perform first-stage paging and second-stage paging by differently adjusting the ratio of base stations used in each of the first service type (iPRS) and the second service type (iPRD).

Total Signaling (TS) for multi-step paging can be calculated as the sum of the number of signals required for each paging step, for example, as shown in Equation 1 below.

Here, Ni represents the paging signal generated in paging step i, and can be calculated by substituting the number of target base stations used in paging step i according to the selected paging service type. Additionally, S may represent the paging success rate in paging step i.

Likewise, Total Delay (TD) can be calculated by adding the delay occurring in each paging step i. For example, if paging is successful, TD corresponds to the time (Tresp) required to receive a page response from the user terminal, and if paging is unsuccessful, TD is the expiration time (expiry) of the paging timer (e.g. T3513) in paging step i. time) may correspond to the T3513i. TD can be calculated, for example, as in Equation 2 below.

FIG. 10 is a diagram illustrating operations in an offline training process and an online prediction process between a data analysis device and a mobility management device according to an embodiment.

The third-generation partnership project's (3GPP) recent 5G specifications add a Network Data Analytics Function (NWDAF) to facilitate artificial intelligence (AI)-based optimization of network functions such as paging. It can be included.

A data analysis device including the second data analysis device 1001 and the first data analysis devices 1003 according to an embodiment may correspond to a device capable of performing a network data analysis function.

Referring to FIG. 10, the offline training process 1007 of the prediction model 1050 (e.g., the prediction model 700 of FIG. 7) according to an embodiment and the base station list predicted by the trained prediction model 1050 ( The structure 1000 of an integrated platform as a service (iPaaS) that performs an online prediction process 1009 using 1006) (e.g., base station list 750 in FIG. 7) is shown.

Integration Platform as a Service (iPaaS) according to one embodiment can extend the centralized structure to distributed NWDAF instances using, for example, serverless computing.

The offline training process 1007 may involve, for example, data collection, clustering, and/or predictive model training, and may utilize significant computational resources. Since the offline training process 1007 may be performed at specific times, managing idle resources while on standby may be required. Serverless computing can reduce the occurrence of idle resources because resources are allocated according to the requests of computing devices, such as, for example, data analysis devices (e.g., data analysis device 1100 in FIG. 11) while functions are operating. there is. In one embodiment, resources can be managed efficiently by distributing offline training as a service to a cloud server (or the second data analysis device 1001) using serverless computing.

In contrast, the online prediction process 1009 uses fewer computational resources than the offline training process 1007, but can be performed more frequently. The online prediction process 1009 reports prediction results to mobility management devices 1065 (e.g., AMF in FIG. 1) in real time due to the high transmission time between the 5G core network and the cloud server (or the second data analysis device 1001). 130), the mobility management device 310 of FIG. 3, the mobility management device 805 of FIG. 8, and/or the mobility management device 1200 of FIG. 12).

The integrated platform as a service (iPaaS) can transmit the trained prediction model 1050 to the first data analysis devices 1060 distributed within the network to increase the response time of real-time prediction and facilitate load balancing. As a result, the integrated platform as a service (iPaaS) structure can efficiently manage resources by eliminating idle resources and distributing the load among data analysis devices.

The offline training process 1007 may be performed using a serverless computing structure and a cloud server by one second data analysis device 1001 and first data analysis devices 1003 distributed within the network. One second data analysis device 1001 may be, for example, a cloud server or a separate device that performs the NWDAF function. In addition, the online prediction process 1009 is performed on the mobility management devices 1065 using the base station list 1006 predicted by the trained prediction model 1050 stored in the distributed first data analysis devices 1060. It can be performed by Here, the distributed first data analysis devices 1060 may be the same devices as the first data analysis devices 1003 and may store the trained prediction model 1050.

The first data analysis devices 1003 may collect historical mobility data of the user terminal(s) in the active mode through the mobility management devices 1005 to determine the location of the user terminal. In terms of structure, data analysis devices can be directly connected to all 5G core networks to collect data and deliver analyzed results. At this time, if only one logically centralized data analysis device is used, the performance of all artificial intelligence-based services may be reduced by increasing communication delay with the 5G core network. Additionally, deep neural networks and data analysis models require high computational resources, and since data analysis devices host not one but multiple artificial intelligence models for different network functions, many computational resources may be utilized. For example, in the current cloud-native network function (CNF) implementation of the 5G core network, resources are continuously allocated even when the user terminal is idle, which may waste the resources of large-scale data analysis devices.

In one embodiment, the serverless computing paradigm can be applied to the data analysis device and the location-based distributed NWDAF structure can be used to prevent the generation of idle resources and distribute the load.

Operations 1010 to 1040 may be performed in the offline training process 1007, and operation 1070 may be performed in the online prediction process 1009.

In operation 1010, mobility management devices 1005, such as AMF devices, transmit mobility data that samples the movement path of the user terminal(s) at fixed intervals to the first data analysis devices 1003 distributed based on location. You can.

In operation 1020, each of the distributed first data analysis devices 1003 may periodically transmit mobility data to the data collection function of the second data analysis device (or cloud server) 1001 to reduce the computational load. Depending on the embodiment, the second data analysis device (or cloud server) 1001 may group user terminal(s) into clusters according to time and mobility patterns, for example, to ensure differentiated services for each cluster.

In operation 1030, the second data analysis device (or cloud server) 1001 may preprocess the mobility data and perform offline training of the prediction model 1050 using the preprocessed mobility data. Offline training of the prediction model 1050 may be performed using, for example, the movement path S, the elapsed time of the user terminal in idle mode ('second elapsed time') (t), and/or the identification information of the user terminal (or It may be performed by inputting two or more or a combination of (c) (identification information of the cluster to which the user terminal belongs).

The prediction model 1050 outputs a base station list 1006 including the probability of the user terminal being located for each base station from the output of the stacked GRUs, for example, by using the stacked GRUs and a softmax activation function. may include fully connected layers.

Stacked GRUs belong to a class of recurrent models for time series data and can achieve similar performance while requiring less computational space than LSTMs. Stacked GRUs may be composed of three layers each consisting of n GRU cells. The first layer of the stacked GRUs may be inputted with the user terminal's previous movement path (S) and the second elapsed time (t) in idle mode, or the user terminal's previous movement path (S) and the second elapsed time (t) in idle mode. 2 Elapsed time (t) and identification information (c) of the cluster to which the user terminal belongs may be input.

Here, since the second elapsed time (t) and the cluster identification information (c) are not time series data, the prediction model 1050 can be trained regardless of the order by providing the same value to all cells of the input layer. Among the stacked GRUs, the second and third layers serve as hidden layers and can contribute to learning the basic mobility pattern of the input sequence. The output of GRU cells can be transmitted starting from the first cell of the input layer to the next cell of the same layer and all of the following layers. This sequential operation can continue until the last GRU cell of the top layer, where the result becomes the final output of the stacked GRU.

Additionally, the output layer of the prediction model 1050 can calculate the probability of a user terminal being located for each base station through fully connected layers. Fully connected layers can receive the output of the stacked GRUs and calculate the probability that a user terminal exists for each j base stations using the softmax activation function. At this time, the sum of all probabilities that the user terminal exists may always be 1.

One second data analysis device 1001 or first data analysis devices 1003 distributed within the network sorts the user terminal location probabilities for each base station output through fully connected layers, for example, in descending order. , a base station list can be created by selecting base stations corresponding to a certain percentage of the sorted probabilities (e.g., top 20%). At this time, the top 20% may be determined through experimentation to balance prediction accuracy and signaling overhead, but are not necessarily limited thereto.

In operation 1040, the second data analysis device (or cloud server) 1001 executes the trained prediction model 1050 together with the identification information of the user terminal (or the identification information of the cluster to which the user terminal belongs) through an online prediction process 1009. It may be propagated to the distributed first data analysis devices 1060 for performance of.

In operation 1070 of the online prediction process 1009, the distributed first data analysis devices 1060 may transmit the base station list 1006 predicted by the prediction model 1050 to the mobility management devices 1065. Mobility management devices 1065 may perform multi-step paging using base stations included in the base station list 1006 according to the paging service type of the user terminal.

Figure 11 is a block diagram of a data analysis device according to one embodiment. Referring to FIG. 11, the data analysis device 1100 according to an embodiment includes first data analysis devices distributed within a network (e.g., the first data analysis devices 1003 and 1060 of FIG. 10) and one centralized device. may include a second data analysis device (eg, the second data analysis device 1001 of FIG. 10).

The data analysis device 1100 may implement the operations of the first data analysis devices distributed within the network and the centralized second data analysis device by one cloud server, or by separate data analysis devices. It may be possible.

Data analysis device 1100 may include a communication interface 1110, a processor 1130, and memory 1150. The communication interface 1110, processor 1130, and memory 1150 may be connected to each other through a communication bus 1105.

The communication interface 1110 is connected to mobility management devices in the network (e.g., AMF 130 in FIG. 1, mobility management device 310 in FIG. 3, mobility management device 805 in FIG. 8, mobility management device in FIG. 10 ( 1005, 1065), and/or mobility management device 1200 of FIG. 12) from a user terminal (e.g., user terminal 110 of FIG. 1, user terminals 250, 260 of FIG. 2, and/or the electronic device of FIG. 13). Mobility data of the device 1301 may be collected. The communication interface 1110 displays a list of base stations (e.g., the base station list 750 of FIG. 7) predicted by a trained prediction model (e.g., the prediction model 700 of FIG. 7, and/or the prediction model 1050 of FIG. 10). , and/or the base station list 1006 of FIG. 10) and identification information of the user terminal may be transmitted to the mobility management device.

The processor 1130 preprocesses the mobility data collected through the communication interface 1110 and applies the preprocessed mobility data to a neural network-based prediction model stored in the memory 1150, so that the prediction model predicts the user terminal will move. It can be trained to predict base stations serving the target location. Additionally, the processor 1130 may transmit the base station list predicted by the trained prediction model and the identification information of the user terminal to the mobility management device through the communication interface 1110.

The memory 1150 may store a neural network-based prediction model.

Additionally, the processor 1130 may execute a program and control the data analysis device 1100. Program code executed by the processor 1130 may be stored in the memory 1150.

The memory 1150 may store information received from the communication interface 1110. The memory 1150 stores executable instructions to be executed by the processor 1130. Additionally, the memory 1150 can store various information generated during processing by the processor 1130. In addition, the memory 1150 can store various data and programs. Memory 1150 may include volatile memory or non-volatile memory. The memory 1150 may be equipped with a high-capacity storage medium such as a hard disk to store various data.

Additionally, the processor 1130 may perform at least one method or a technique corresponding to at least one method described above with reference to FIGS. 1 to 10 . The processor 1130 may be a device or server implemented in hardware that has a circuit with a physical structure for executing desired operations. For example, the intended operations may include code or instructions included in the program. For example, the data analysis device 1100 implemented in hardware includes a microprocessor, a central processing unit (CPU), a graphic processing unit (GPU), a processor core, It may include a multi-core processor, multiprocessor, application-specific integrated circuit (ASIC), field programmable gate array (FPGA), and/or neural processing unit (NPU).

Figure 12 is a block diagram of a mobility management device according to one embodiment. Referring to FIG. 12, a mobility management device 1200 according to an embodiment (e.g., AMF 130 in FIG. 1, mobility management device 310 in FIG. 3, mobility management device 805 in FIG. 8, and/ Alternatively, the mobility management devices 1005 and 1065 of FIG. 10 may include a communication interface 1210, a processor 1230, and a memory 1250. The communication interface 1210, processor 1230, and memory 1250 may be connected to each other through a communication bus 1205.

The communication interface 1210 provides a user terminal (e.g., the user terminal 110 of FIG. 1, FIG. 2) to the first data analysis device (e.g., the first data analysis devices 1003 and 1060 of FIG. 10) distributed within the network. Mobility data of the user terminals 250 and 260 and/or the electronic device 1301 of FIG. 13 may be transmitted. The first data analysis device may include a neural network-based trained (eg, prediction model 700 of FIG. 7 and/or prediction model 1050 of FIG. 10). The communication interface 1210 includes a base station list (e.g., the base station list in FIG. 7) that includes base stations serving the target location where the user terminal is expected to move, predicted by the prediction model by the first data analysis device based on the mobility data. 750, and/or base station list 1006 of FIG. 10) may be received.

The processor 1230 determines a target base station that performs paging for each stage of multi-stage paging among base stations (e.g., 5G RAN in FIG. 1, base stations 215, 225, 235, 245 in FIG. 2, and/or base station 330 in FIG. 3). You can. The processor 1230 may perform multi-step paging for the user terminal by the target base station.

Additionally, the processor 1230 may execute a program and control the mobility management device 1200. Program code executed by the processor 1230 may be stored in the memory 1250.

The memory 1250 may store information received from the communication interface 1210. The memory 1250 stores executable instructions to be executed by the processor 1230. Additionally, the memory 1250 can store various information generated during processing by the processor 1230. In addition, the memory 1250 can store various data and programs. Memory 1250 may include volatile memory or non-volatile memory. The memory 1250 may be equipped with a high-capacity storage medium such as a hard disk to store various data.

Additionally, the processor 1230 may perform at least one method or a technique corresponding to at least one method described above with reference to FIGS. 1 to 10 . The processor 1230 may be a communication device implemented in hardware that has a circuit with a physical structure for executing desired operations. For example, the intended operations may include code or instructions included in the program. For example, the mobility management device 1200 implemented in hardware may include a microprocessor, a central processing unit (CPU), a graphics processing unit (GPU), a processor core, and a multi-core processor. It may include a processor, a multiprocessor, an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), and/or a neural processing unit (NPU).

13 is a block diagram of an electronic device in a network environment according to various embodiments. Referring to FIG. 13, a user terminal (e.g., the user terminal 110 of FIG. 1, the user terminals 250 and 260 of FIG. 2, and/or the electronic device 1301 of FIG. 13) in the network environment 1300. The exemplary electronic device 1301 communicates with the electronic device 1302 through a first network 1398 (e.g., a short-range wireless communication network) or through a second network 1399 (e.g., a long-range wireless communication network). It may communicate with at least one of the electronic device 1304 or the server 1308. According to one embodiment, the electronic device 1301 may communicate with the electronic device 1304 through the server 1308. According to one embodiment, the electronic device 1301 includes a processor 1220, a memory 1230, an input module 1250, an audio output module 1355, a display module 1260, an audio module 1370, and a sensor module ( 1376), interface 1377, connection terminal 1378, haptic module 1379, camera module 1380, power management module 1388, battery 1389, communication module 1390, subscriber identification module 1396. , or may include an antenna module 1397. In some embodiments, at least one of these components (eg, the connection terminal 1378) may be omitted or one or more other components may be added to the electronic device 1301. In some embodiments, some of these components (e.g., sensor module 1376, camera module 1380, or antenna module 1397) are integrated into one component (e.g., display module 1260). It can be.

The processor 1220, for example, executes software (e.g., program 1240) to operate at least one other component (e.g., hardware or software component) of the electronic device 1301 connected to the processor 1220. It can be controlled and various data processing or calculations can be performed. According to one embodiment, as at least part of data processing or computation, the processor 1220 stores commands or data received from another component (e.g., sensor module 1376 or communication module 1390) in volatile memory 1332. The commands or data stored in the volatile memory 1332 can be processed, and the resulting data can be stored in the non-volatile memory 1334. According to one embodiment, the processor 1220 includes a main processor 1321 (e.g., a central processing unit or an application processor) or an auxiliary processor 1323 that can operate independently or together (e.g., a graphics processing unit, a neural network processing unit ( It may include a neural processing unit (NPU), an image signal processor, a sensor hub processor, or a communication processor). For example, if the electronic device 1301 includes a main processor 1321 and a auxiliary processor 1323, the auxiliary processor 1323 may be set to use lower power than the main processor 1321 or be specialized for a designated function. You can. The auxiliary processor 1323 may be implemented separately from the main processor 1321 or as part of it.

The auxiliary processor 1323 may, for example, act on behalf of the main processor 1321 while the main processor 1321 is in an inactive (e.g., sleep) state, or while the main processor 1321 is in an active (e.g., application execution) state. ), together with the main processor 1321, at least one of the components of the electronic device 1301 (e.g., the display module 1260, the sensor module 1376, or the communication module 1390) At least some of the functions or states related to can be controlled. According to one embodiment, coprocessor 1323 (e.g., image signal processor or communication processor) may be implemented as part of another functionally related component (e.g., camera module 1380 or communication module 1390). there is. According to one embodiment, the auxiliary processor 1323 (eg, neural network processing unit) may include a hardware structure specialized for processing artificial intelligence models. Artificial intelligence models can be created through machine learning. For example, such learning may be performed in the electronic device 1301 itself on which the artificial intelligence model is performed, or may be performed through a separate server (e.g., server 1308). Learning algorithms may include, for example, supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning, but It is not limited. An artificial intelligence model may include multiple artificial neural network layers. Artificial neural networks include deep neural network (DNN), convolutional neural network (CNN), recurrent neural network (RNN), restricted boltzmann machine (RBM), belief deep network (DBN), bidirectional recurrent deep neural network (BRDNN), It may be one of deep Q-networks or a combination of two or more of the above, but is not limited to the examples described above. In addition to hardware structures, artificial intelligence models may additionally or alternatively include software structures.

The memory 1230 may store various data used by at least one component (eg, the processor 1220 or the sensor module 1376) of the electronic device 1301. Data may include, for example, input data or output data for software (e.g., program 1240) and instructions related thereto. Memory 1230 may include volatile memory 1332 or non-volatile memory 1334.

The program 1240 may be stored as software in the memory 1230 and may include, for example, an operating system 1342, middleware 1344, or application 1346.

The input module 1250 may receive commands or data to be used in a component of the electronic device 1301 (e.g., the processor 1220) from outside the electronic device 1301 (e.g., a user). The input module 1250 may include, for example, a microphone, mouse, keyboard, keys (eg, buttons), or digital pen (eg, stylus pen).

The sound output module 1355 may output sound signals to the outside of the electronic device 1301. The sound output module 1355 may include, for example, a speaker or receiver. Speakers can be used for general purposes such as multimedia playback or recording playback. The receiver can be used to receive incoming calls. According to one embodiment, the receiver may be implemented separately from the speaker or as part of it.

The display module 1260 can visually provide information to the outside of the electronic device 1301 (eg, a user). The display module 1260 may include, for example, a display, a hologram device, or a projector, and a control circuit for controlling the device. According to one embodiment, the display module 1260 may include a touch sensor configured to detect a touch, or a pressure sensor configured to measure the intensity of force generated by the touch.

The audio module 1370 can convert sound into an electrical signal or, conversely, convert an electrical signal into sound. According to one embodiment, the audio module 1370 acquires sound through the input module 1250, the sound output module 1355, or an external electronic device (e.g., directly or wirelessly connected to the electronic device 1301). Sound may be output through an electronic device 1302 (e.g., speaker or headphone).

The sensor module 1376 detects the operating state (e.g., power or temperature) of the electronic device 1301 or the external environmental state (e.g., user state) and generates an electrical signal or data value corresponding to the detected state. can do. According to one embodiment, the sensor module 1376 includes, for example, a gesture sensor, a gyro sensor, an air pressure sensor, a magnetic sensor, an acceleration sensor, a grip sensor, a proximity sensor, a color sensor, an IR (infrared) sensor, a biometric sensor, It may include a temperature sensor, humidity sensor, or light sensor.

The interface 1377 may support one or more designated protocols that can be used to directly or wirelessly connect the electronic device 1301 to an external electronic device (e.g., the electronic device 1302). According to one embodiment, the interface 1377 may include, for example, a high definition multimedia interface (HDMI), a universal serial bus (USB) interface, an SD card interface, or an audio interface.

The connection terminal 1378 may include a connector through which the electronic device 1301 can be physically connected to an external electronic device (eg, the electronic device 1302). According to one embodiment, the connection terminal 1378 may include, for example, an HDMI connector, a USB connector, an SD card connector, or an audio connector (eg, a headphone connector).

The haptic module 1379 can convert electrical signals into mechanical stimulation (e.g., vibration or movement) or electrical stimulation that the user can perceive through tactile or kinesthetic senses. According to one embodiment, the haptic module 1379 may include, for example, a motor, a piezoelectric element, or an electrical stimulation device.

The camera module 1380 can capture still images and moving images. According to one embodiment, the camera module 1380 may include one or more lenses, image sensors, image signal processors, or flashes.

The power management module 1388 can manage power supplied to the electronic device 1301. According to one embodiment, the power management module 1388 may be implemented as at least a part of, for example, a power management integrated circuit (PMIC).

The battery 1389 may supply power to at least one component of the electronic device 1301. According to one embodiment, the battery 1389 may include, for example, a non-rechargeable primary cell, a rechargeable secondary cell, or a fuel cell.

Communication module 1390 provides a direct (e.g., wired) communication channel or wireless communication channel between the electronic device 1301 and an external electronic device (e.g., electronic device 1302, electronic device 1304, or server 1308). It can support establishment and communication through established communication channels. Communication module 1390 operates independently of processor 1220 (e.g., an application processor) and may include one or more communication processors that support direct (e.g., wired) communication or wireless communication. According to one embodiment, the communication module 1390 may be a wireless communication module 1392 (e.g., a cellular communication module, a short-range wireless communication module, or a global navigation satellite system (GNSS) communication module) or a wired communication module 1394 (e.g., : LAN (local area network) communication module, or power line communication module) may be included. Among these communication modules, the corresponding communication module is a first network 1398 (e.g., a short-range communication network such as Bluetooth, wireless fidelity (WiFi) direct, or infrared data association (IrDA)) or a second network 1399 (e.g., legacy It may communicate with an external electronic device 1304 through a telecommunication network such as a cellular network, a 5G network, a next-generation communication network, the Internet, or a computer network (e.g., LAN or WAN). These various types of communication modules may be integrated into one component (e.g., a single chip) or may be implemented as a plurality of separate components (e.g., multiple chips). The wireless communication module 1392 uses subscriber information (e.g., International Mobile Subscriber Identifier (IMSI)) stored in the subscriber identification module 1396 to communicate within a communication network such as the first network 1398 or the second network 1399. The electronic device 1301 can be confirmed or authenticated.

The wireless communication module 1392 may support 5G networks and next-generation communication technologies after 4G networks, for example, NR access technology (new radio access technology). NR access technology provides high-speed transmission of high-capacity data (eMBB (enhanced mobile broadband)), minimization of terminal power and access to multiple terminals (mMTC (massive machine type communications)), or high reliability and low latency (URLLC (ultra-reliable and low latency). -latency communications)) can be supported. The wireless communication module 1392 may support high frequency bands (e.g., mmWave bands), for example, to achieve high data rates. The wireless communication module 1392 uses various technologies to secure performance in high frequency bands, such as beamforming, massive MIMO (multiple-input and multiple-output), and full-dimensional multiplexing. It can support technologies such as input/output (FD-MIMO: full dimensional MIMO), array antenna, analog beam-forming, or large scale antenna. The wireless communication module 1392 may support various requirements specified in the electronic device 1301, an external electronic device (e.g., electronic device 1304), or a network system (e.g., second network 1399). According to one embodiment, the wireless communication module 1392 supports peak data rate (e.g., 20 Gbps or more) for realizing eMBB, loss coverage (e.g., 164 dB or less) for realizing mmTC, or U-plane latency (e.g., 164 dB or less) for realizing URLLC. Example: Downlink (DL) and uplink (UL) each of 0.5 ms or less, or round trip 1 ms or less) can be supported.

The antenna module 1397 may transmit or receive signals or power to or from the outside (e.g., an external electronic device). According to one embodiment, the antenna module 1397 may include an antenna including a radiator made of a conductor or a conductive pattern formed on a substrate (eg, PCB). According to one embodiment, the antenna module 1397 may include a plurality of antennas (eg, an array antenna). In this case, at least one antenna suitable for a communication method used in a communication network such as the first network 1398 or the second network 1399 is connected to the plurality of antennas by, for example, the communication module 1390. can be selected Signals or power may be transmitted or received between the communication module 1390 and an external electronic device through the at least one selected antenna. According to some embodiments, in addition to the radiator, other components (eg, radio frequency integrated circuit (RFIC)) may be additionally formed as part of the antenna module 1397.

According to various embodiments, antenna module 1397 may form a mmWave antenna module. According to one embodiment, a mmWave antenna module includes: a printed circuit board, an RFIC disposed on or adjacent to a first side (e.g., bottom side) of the printed circuit board and capable of supporting a designated high frequency band (e.g., mmWave band); And a plurality of antennas (e.g., array antennas) disposed on or adjacent to the second side (e.g., top or side) of the printed circuit board and capable of transmitting or receiving signals in the designated high frequency band. can do.

At least some of the components are connected to each other through a communication method between peripheral devices (e.g., bus, general purpose input and output (GPIO), serial peripheral interface (SPI), or mobile industry processor interface (MIPI)) and signal ( (e.g. commands or data) can be exchanged with each other.

According to one embodiment, commands or data may be transmitted or received between the electronic device 1301 and the external electronic device 1304 through the server 1308 connected to the second network 1399. Each of the external electronic devices 1302 or 1304 may be of the same or different type as the electronic device 1301. According to one embodiment, all or part of the operations performed in the electronic device 1301 may be executed in one or more of the external electronic devices 1302, 1304, or 1308. For example, when the electronic device 1301 needs to perform a certain function or service automatically or in response to a request from a user or another device, the electronic device 1301 does not execute the function or service on its own. Alternatively, or additionally, one or more external electronic devices may be requested to perform at least part of the function or service. One or more external electronic devices that have received the request may execute at least part of the requested function or service, or an additional function or service related to the request, and transmit the result of the execution to the electronic device 1301. The electronic device 1301 may process the result as is or additionally and provide it as at least part of a response to the request. For this purpose, for example, cloud computing, distributed computing, mobile edge computing (MEC), or client-server computing technology can be used. The electronic device 1301 may provide an ultra-low latency service using, for example, distributed computing or mobile edge computing. In another embodiment, the external electronic device 1304 may include an Internet of Things (IoT) device. Server 1308 may be an intelligent server using machine learning and/or neural networks. According to one embodiment, the external electronic device 1304 or server 1308 may be included in the second network 1399. The electronic device 1301 may be applied to intelligent services (e.g., smart home, smart city, smart car, or healthcare) based on 5G communication technology and IoT-related technology.

According to one embodiment, a method of operating a data analysis device including first data analysis devices 1003 and 1060 distributed in a network and a centralized second data analysis device 1001 includes the first data analysis devices ( Operation 510, where 1003, 1060) collects mobility data of the user terminal (110, 250, 260, 1301) from mobility management devices (130, 310, 805, 1005, 1065, 1200) in the network, the second In operation 520, the data analysis device 1001 preprocesses the mobility data, and the second data analysis device 1001 applies the preprocessed mobility data to neural network-based prediction models 700 and 1050 to make the prediction. Operation 530 of training a model (700, 1050) to predict base stations serving a target location where the user terminal (110, 250, 260, 1301) is expected to move, and the first data analysis devices (1003, 1060) ) may include an operation 540 of transmitting the trained prediction model (700, 1050) and the identification information of the user terminal (110, 250, 260, 1301).

According to one embodiment, the mobility data includes location information that the user terminal (110, 250, 260, 1301) has moved in an activation mode in which the user terminal (110, 250, 260, 1301) moves in the activation mode. It may include any one or a combination of a first elapsed time in the mode and a second elapsed time in an idle mode in which the user terminal (110, 250, 260, 1301) does not perform the communication. .

According to one embodiment, the location information corresponds to identification information of the user terminal (110, 250, 260, 1301) and a first location from which the user terminal (110, 250, 260, 1301) started in the activation mode. Information about the first base station, and information about the second base station corresponding to the second location where the user terminal (110, 250, 260, 1301) moves from the first location in the activation mode, or It may include combinations of these.

According to one embodiment, the operation of preprocessing the mobility data includes dividing the mobility data by identification information of the user terminals 110, 250, 260, and 1301 to generate one continuous sequence, the continuous one An operation of dividing the sequence into a plurality of unit sequences including location information corresponding to a fixed time unit, and the user terminal (110, 250, 260, 1301) during the first elapsed time in the activation mode The movement path of the user terminal (110, 250, 260, 1301) and the user terminal (110, 250, 260, 260, 260, 1301), the operation of configuring the plurality of unit sequences into a batch for learning and evaluating the prediction models 700 and 1050 may be included.

According to one embodiment, the operation of configuring the plurality of unit sequences into the arrangement involves using the first unit sequences belonging to the first elapsed time among the plurality of unit sequences for training the prediction model (700, 1050). An operation of setting as an input, an operation of setting second unit sequences belonging to the second elapsed time among the plurality of unit sequences as labels for evaluation of the prediction models (700, 1050), and the user terminal Identification information of (110, 250, 260, 1301), the second elapsed time, the movement path of the user terminal (110, 250, 260, 1301) during the first elapsed time, and after the second elapsed time It may include an operation of configuring information including the target location of the user terminal (110, 250, 260, 1301) into the arrangement.

According to one embodiment, the training operation uses mobility data sampled over a fixed time interval and a first elapsed time when the user terminal (110, 250, 260, 1301) is in an active mode. This may include training the prediction models 700 and 1050 to predict the base stations serving the target location.

According to one embodiment, the prediction model (700, 1050) is a stacked deep neural network (710) that receives the preprocessed mobility data and learns spatiotemporal characteristics of the mobility of the user terminal (110, 250, 260, 1301). ), and a fully connected layer that outputs a base station list (750, 1006) including the probability that the user terminal (110, 250, 260, 1301) is located for each base station from the output of the stacked deep neural network (710). may include fully connected layers 730.

According to one embodiment, the stacked deep neural network 710 includes an input layer composed of deep neural network cells, and hidden layers, and the fixed time of the user terminals 110, 250, 260, and 1301 is applied to the input layer. As previous location paths according to the interval and a second elapsed time elapsed when the user terminal (110, 250, 260, 1301) is in idle mode are applied, the hidden layers are applied to the user based on the previous location paths. The terminal (110, 250, 260, 1301) may train the target location that is expected to move in the second elapsed time.

According to one embodiment, the data analysis device sorts the probability that the user terminal (110, 250, 260, 1301) is located for each of the base stations output through the fully connected layers 730, and the sorted The base station list (750, 1006) containing base stations corresponding to a certain percentage of probabilities can be output.

According to one embodiment, the method of operating the mobility management devices (130, 310, 805, 1005, 1065, 1200) includes: a first data analysis device distributed within a network - the first data analysis device is a trained prediction model based on a neural network; Including (700, 1050) - operation 810 of transmitting mobility data of the user terminal (110, 250, 260, 1301), wherein the first data analysis device transmits the mobility data to the prediction model (700, 1050). Operation 820, receiving a base station list (750, 1006) including base stations serving the predicted target location to which the user terminal (110, 250, 260, 1301) is expected to move, the base station list (750, Operation 830 of determining a target base station that performs paging for each stage of multi-level paging among the base stations included in 1006), and performing the multi-level paging for the user terminal (110, 250, 260, 1301) by the target base station. It may include the operation 840 to be performed.

According to one embodiment, the operation of determining the target base station is performed when it is determined that the user terminal (110, 250, 260, 1301) is not present in the cell of the last known base station in the activation mode. , 250, 260, 1301), an operation of determining the target base station among the base stations included in the base station list (750, 1006) may be included.

According to one embodiment, the operation of determining the target base station involves determining that the paging service type corresponding to the user terminal (110, 250, 260, 1301) is a first service type (iPRS) for reducing signaling and reducing delay. An operation of determining which paging service type is a second service type (iPRD) for a second service type (iPRD), and a first target base station used for first-stage paging and a first target base station used for second-stage paging among the base stations according to the determined paging service type. 2 It may include an operation to adjust the ratio of the target base station.

According to one embodiment, the operation of determining which paging service type is based on one of the service type and charging policy corresponding to the user terminal (110, 250, 260, and 1301), 250, 260, 1301) may include determining which paging service type corresponds to the paging service type.

According to one embodiment, the operation of adjusting the ratio of the first target base station and the second target base station is performed by the user terminal (110, 250, 260, 1301) when the determined paging service type is determined to be the first service type. ) may include determining the last known base station in this activation mode as the first target base station, and determining the base stations included in the base station list (750, 1006) as the second target base station.

According to one embodiment, the operation of adjusting the ratio of the first target base station and the second target base station is performed when the determined paging service type is determined to be the second service type, included in the base station list (750, 1006). An operation of determining a number of base stations corresponding to a first ratio among the base stations as the first target base stations, and a second target base station remaining excluding the first ratio among the base stations included in the base station list (750, 1006). It may include an operation of determining a number of base stations corresponding to the ratio as the second target base stations.

According to one embodiment, the operation of performing the multi-stage paging includes performing the first stage paging by the first target base station, and performing the second stage paging by the second target base station. can do.

According to one embodiment, the mobility data includes location information that the user terminal (110, 250, 260, 1301) has moved in an activation mode in which the user terminal (110, 250, 260, 1301) moves in the activation mode. It may include any one or a combination of a first elapsed time in the mode and a second elapsed time in an idle mode in which the user terminal (110, 250, 260, 1301) does not perform the communication. .

According to one embodiment, the data analysis device 1100, which includes first data analysis devices 1003 and 1060 distributed within the network and a centralized second data analysis device 1001, is connected to the mobility management device 130 within the network. , 310, 805, 1005, 1065, 1200, a communication interface 1110 that collects mobility data of the user terminals 110, 250, 260, 1301, and pre-processes the mobility data, and By applying it to the neural network-based prediction model (700, 1050), the prediction model (700, 1050) is trained to predict base stations serving the target location where the user terminal (110, 250, 260, 1301) is expected to move. It includes a processor 1130, and the communication interface 1110 includes a base station list 750, 1006 predicted by the trained prediction model 700, 1050 and the user terminal 110, 250, 260, 1301. Identification information may be transmitted to the mobility management devices 130, 310, 805, 1005, 1065, and 1200.

According to one embodiment, the mobility management devices (130, 310, 805, 1005, 1065, 1200) are a first data analysis device distributed within a network - the first data analysis device is a neural network-based trained prediction model (700, 1050) - transmitting mobility data of a user terminal (110, 250, 260, 1301), and predicting by the first data analysis device by the prediction model (700, 1050) based on the mobility data. , a communication interface 1210 that receives a base station list 750, 1006 including base stations serving target locations to which the user terminal 110, 250, 260, 1301 is expected to move, and multi-level paging among the base stations. It may include a processor 1230 that determines a target base station that performs paging for each step, and performs the multi-step paging for the user terminals 110, 250, 260, and 1301 by the target base station.

Claims (20)

In a method of operating a data analysis device including first data analysis devices distributed in a network and a second data analysis device centralized,
collecting, by the first data analysis devices, mobility data of a user terminal from mobility management devices in the network;
Preprocessing, by the second data analysis device, the mobility data;
an operation, by the second data analysis device, of applying the preprocessed mobility data to a neural network-based prediction model to train the prediction model to predict base stations serving a target location where the user terminal is expected to move; and
An operation of transmitting the trained prediction model and identification information of the user terminal to the first data analysis devices.
A method of operating a data analysis device, including. According to paragraph 1,
The mobility data is
Location information that the user terminal has moved in the active mode in which the user terminal performs communication, a first elapsed time that the user terminal has elapsed in the active mode, and a second elapsed time in the idle mode in which the user terminal has not performed the communication. A method of operating a data analysis device comprising any one or a combination of time. According to paragraph 2,
The location information is
Identification information of the user terminal, information on a first base station corresponding to the first location from which the user terminal departed in the activation mode, and a second location that the user terminal arrived at by moving from the first location in the activation mode Any one or a combination of information about the second base station corresponding to
A method of operating a data analysis device, including. According to paragraph 1,
The operation of preprocessing the mobility data is
An operation of dividing the mobility data according to identification information of the user terminal to generate one continuous sequence;
dividing the single continuous sequence into a plurality of unit sequences including location information corresponding to a fixed time unit; and
Based on the movement path of the user terminal during a first elapsed time when the user terminal is in an active mode and the target location of the user terminal during a second elapsed time when the user terminal is in an idle mode, the plurality of An operation of organizing unit sequences into a batch for learning and evaluating the prediction model.
A method of operating a data analysis device, including. According to paragraph 4,
The operation of configuring the plurality of unit sequences into the arrangement is
Setting first unit sequences belonging to the first elapsed time among the plurality of unit sequences as input for training the prediction model;
Setting second unit sequences belonging to the second elapsed time among the plurality of unit sequences as labels for evaluating the prediction model; and
Information including the identification information of the user terminal, the second elapsed time, the movement path of the user terminal during the first elapsed time, and the target location of the user terminal after the second elapsed time are arranged in the arrangement. Constructing Actions
A method of operating a data analysis device, including. According to paragraph 1,
The training movement is
An operation of training the prediction model to predict the base stations serving the target location using mobility data sampled over a fixed time interval and a first elapsed time during which the user terminal is in an active mode.
A method of operating a data analysis device, including. According to paragraph 1,
The prediction model is
A layered deep neural network that receives the preprocessed mobility data and learns spatiotemporal characteristics of the user terminal's mobility; and
Fully connected layers that output a base station list including the probability of the user terminal being located for each base station from the output of the stacked deep neural network.
A method of operating a data analysis device, including. In clause 7,
The layered deep neural network is
Input layer composed of deep neural network cells, and hidden layers
Including,
As previous location paths according to a fixed time interval of the user terminal and a second elapsed time elapsed when the user terminal is in idle mode are applied to the input layer, the hidden layers are based on the previous location paths. A method of operating a data analysis device, training the target location to which the user terminal is expected to move in the second elapsed time. In clause 7,
The data analysis device is
A data analysis device that sorts the probability of the user terminal being located for each of the base stations output through the fully connected layers and outputs the base station list including base stations corresponding to a certain percentage of the sorted probabilities. How it works. An operation of transmitting mobility data of a user terminal to a first data analysis device distributed within a network, wherein the first data analysis device includes a trained prediction model based on a neural network;
Receiving a base station list including base stations serving a target location to which the user terminal is expected to move, predicted by the first data analysis device by applying the mobility data to the prediction model;
An operation of determining a target base station that performs paging for each stage of multi-stage paging among the base stations included in the base station list; and
An operation of performing the multi-step paging for the user terminal by the target base station
A method of operating a mobility management device, including. According to clause 10,
The operation of determining the target base station is
If it is determined that the user terminal is not in the cell of the last known base station in active mode,
An operation of determining the target base station among the base stations included in the base station list, according to the paging service type corresponding to the user terminal.
A method of operating a mobility management device, including. According to clause 11,
The operation of determining the target base station is
determining which paging service type corresponding to the user terminal is a first service type for reducing signaling (iPRS) and a second service type for reducing delay (iPRD); and
An operation of adjusting the ratio of a first target base station used for first-stage paging and a second target base station used for second-stage paging among the base stations according to the determined paging service type.
A method of operating a mobility management device, including. According to clause 12,
The operation of determining which paging service type is
An operation of determining which paging service type is the paging service type corresponding to the user terminal, based on one of a service type and a charging policy corresponding to the user terminal.
A method of operating a mobility management device, including. According to clause 12,
The operation of adjusting the ratio of the first target base station and the second target base station is
When it is determined that the determined paging service type is the first service type,
An operation of the user terminal determining the last known base station in an activation mode as the first target base station; and
An operation of determining the base stations included in the base station list as the second target base station.
A method of operating a mobility management device, including. According to clause 12,
The operation of adjusting the ratio of the first target base station and the second target base station is
When it is determined that the determined paging service type is the second service type,
determining a number of base stations corresponding to a first ratio among the base stations included in the base station list as the first target base stations; and
An operation of determining the number of base stations corresponding to the second ratio of the remaining base stations excluding the first ratio among the base stations included in the base station list as the second target base stations.
A method of operating a mobility management device, including. According to clause 12,
The operation of performing the multi-step paging is
performing the first stage paging by the first target base station; and
An operation of performing the second stage paging by the second target base station
A method of operating a mobility management device, including. According to clause 11,
The mobility data is
Location information that the user terminal has moved in the active mode in which the user terminal performs communication, a first elapsed time that the user terminal has elapsed in the active mode, and a second elapsed time in the idle mode in which the user terminal has not performed the communication. A method of operating a mobility management device, including any one or a combination of time. A computer program combined with hardware and stored in a computer-readable storage medium to execute the method of any one of claims 1 to 17. A data analysis device comprising first data analysis devices distributed within a network and a second data analysis device centralized,
a communication interface that collects mobility data of a user terminal from mobility management devices in the network; and
A processor that preprocesses the mobility data, applies the preprocessed mobility data to a neural network-based prediction model, and trains the prediction model to predict base stations serving the target location where the user terminal is expected to move.
Including,
The communication interface is
A data analysis device that delivers a list of base stations predicted by the trained prediction model and identification information of the user terminal to the mobility management devices. Transmitting mobility data of the user terminal to a first data analysis device distributed within a network, wherein the first data analysis device includes a trained prediction model based on a neural network, and transmitting mobility data of the user terminal to the first data analysis device based on the mobility data. a communication interface that receives a base station list including base stations serving a target location to which the user terminal is expected to move, predicted by the prediction model; and
A processor that determines a target base station that performs paging for each stage of multi-stage paging among the base stations, and performs the multi-stage paging for the user terminal by the target base station.
Mobility management device including.

Images