Communication control device, communication terminal, and communication control method

Abstract

To perform communication control about use of a radio resource to an IoT terminal so as to be an intended communication traffic state.SOLUTION: A communication control device 1 includes a first setting part 10 for setting a distribution of observation data being a set of observation values about a mixed probability model with a use state of each resource block obtained by dividing a radio resource as the observation value of each cluster, a second setting part 11 for setting a parameter of the mixed probability model so as to be a set distribution of observation data, a learning part 12 for performing adversarial learning of a generation model having a generator 121 for generating pseudo use information similar to true use information with the mixed probability model having the set parameter as the true use information about the use state of the resource block and a discriminator 122, and a notification part 14 for notifying a communication terminal 2 of a learned generator 121' constructed by the learning part 12 as communication control information.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a communication control device, a communication terminal, and a communication control method. [Background technology]

[0002] In recent years, IoT devices such as home appliances, electricity, water, and gas smart meters, as well as smartphones and tablets, have become widespread. As the number of IoT devices increases, controlling the traffic in IoT device communications has become an issue. For example, when multiple IoT devices are powered on, traffic can become concentrated, causing burst traffic.

[0003] Therefore, Patent Document 1 discloses a technology for smoothing traffic and reducing congestion in communications between IoT terminals that periodically generate traffic by controlling the delivery time from a time synchronization server to each IoT terminal. However, with the communication control system disclosed in Patent Document 1, it is difficult to similarly control traffic that does not periodically occur. As such, with the communication control system disclosed in Patent Document 1, it is difficult to control the usage of wireless resources to achieve the leveled traffic intended by the network administrator, regardless of whether the traffic occurs periodically or not. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent No. 7472968 Summary of the Invention [Problem to be solved by the invention]

[0005] According to conventional technologies, it has been difficult to control communication regarding the use of wireless resources for IoT terminals so as to achieve the intended communication traffic state.

[0006] The present invention has been made to solve the above-mentioned problems, and aims to perform communication control regarding the use of wireless resources for IoT terminals so that the intended communication traffic state is achieved. [Means for solving the problem]

[0007] In order to solve the above-mentioned problems, the communication control device of the present invention comprises: a first setting unit configured to set a distribution of observation data, which is a set of observation values, for a mixed probability model in which the usage status of each resource block obtained by dividing wireless resources is set as an observation value for each cluster; a second setting unit configured to set parameters of the mixed probability model so that the distribution of the set observation data becomes the set distribution of the observation data; a learning unit configured to perform adversarial learning of a generative model using the mixed probability model having the set parameters as true usage information regarding the usage status of the resource block, a generator that generates pseudo usage information similar to the true usage information, and a discriminator that distinguishes between the pseudo usage information generated by the generator and the true usage information; and a notification unit configured to notify a communication terminal of the learned generator constructed by the learning unit as communication control information.

[0008] In addition, in the communication control device of the present invention, the mixed probability model may be a model that follows a mixed Bernoulli distribution, and the parameters of the mixed probability model may include a mixed ratio that represents the probability that each cluster will generate the observed value, and an average probability that represents the average probability that the resource block will be used in each cluster.

[0009] In the communication control device according to the present invention, the first setting unit may set a probability that the observed value occurs for each cluster.

[0010] In addition, in the communication control device of the present invention, the resource block may be a time slot obtained by dividing the radio resource in the time domain into multiple parts, and the observation value may be a value representing, as binary data, whether or not the communication terminal transmits a signal for each time slot.

[0011] In order to solve the above-mentioned problems, the communication terminal of the present invention comprises a generation unit configured to generate the pseudo-usage information using the learned generator based on the communication control information notified from the above-mentioned communication control device, an identification unit configured to identify resource blocks to be used based on the pseudo-usage information generated by the generation unit, and a communication control unit that instructs the transmission of a signal using the identified resource blocks.

[0012] In order to solve the above-mentioned problems, the communication control method of the present invention includes a first setting step of setting a distribution of observation data, which is a set of observation values, for a mixed probability model in which the usage status of each resource block obtained by dividing wireless resources is used as an observation value for each cluster; a second setting step of setting parameters of the mixed probability model so that the distribution of the set observation data becomes the set distribution of the observation data; a learning step of performing adversarial learning of a generative model using the mixed probability model having the set parameters as true usage information regarding the usage status of the resource block, a generator that generates pseudo-usage information similar to the true usage information, and a discriminator that distinguishes between the pseudo-usage information generated by the generator and the true usage information; and a notification step of notifying a communication terminal of the trained generator constructed in the learning step as communication control information.

[0013] In the communication control method according to the present invention, the first setting step may set a probability that the observed value occurs for each cluster. [Effects of the Invention]

[0014] According to the present invention, adversarial learning of a generative model is performed using a mixture probability model with set parameters as true usage information regarding the usage status of resource blocks, a generator that generates pseudo-usage information similar to the true usage information, and a discriminator that distinguishes between the pseudo-usage information generated by the generator and previous allocation information. As a result, communication control regarding the use of wireless resources can be performed on IoT terminals so that the intended communication traffic state is achieved. [Brief explanation of the drawings]

[0015] [Figure 1] FIG. 1 is a block diagram showing a configuration of a communication control system including a communication control device and a communication terminal according to an embodiment of the present invention. [Figure 2] FIG. 2 is a diagram for explaining an outline of the communication control system according to the present embodiment. [Figure 3] FIG. 3 is a diagram for explaining a learning unit included in the communication control device according to the present embodiment. [Figure 4] FIG. 4 is a diagram for explaining the learning unit included in the communication control device according to the present embodiment. [Figure 5] FIG. 5 is a diagram for explaining a learning unit included in the communication control device according to the present embodiment. [Figure 6] FIG. 6 is a block diagram showing the hardware configuration of the communication control device according to this embodiment. [Figure 7] FIG. 7 is a block diagram showing a hardware configuration of a communication terminal according to this embodiment. [Figure 8] FIG. 8 is a sequence diagram showing the operation of the communication control system according to the present embodiment. [Figure 9] FIG. 9 is a flowchart showing the operation of the communication control device according to the present embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0016] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will now be described in detail with reference to FIGS.

[0017] [Communication control system configuration] First, with reference to FIG. 1, an overview of a communication control system including a communication control device 1 and a communication terminal 2 according to an embodiment of the present invention will be described.

[0018] The communication control system is provided in a mobile communication network such as LTE / 4G, 5G, or 6G. A communication control device 1 and a communication terminal 2 are connected via a network NW such as a WAN or the Internet. The communication control device 1 notifies the communication terminal 2 of communication control information via the network NW. Furthermore, the communication terminal 2 identifies an available time slot based on the notified communication control information, and transmits a signal using the identified time slot.

[0019] The communication terminal 2 is realized by an IoT terminal having a unique IP address, such as a mobile communication terminal such as a smartphone, a tablet computer, a laptop computer, or a sensor device. In this embodiment, there are multiple communication terminals 2. For example, if the communication terminals 2 are K smartphones (K is an integer equal to or greater than 2), they transmit a location registration request signal to a core network (not shown) when they are turned on, at regular intervals, and as the communication terminals 2 move. If K communication terminals 2 simultaneously transmit signals within a certain period of time, traffic will increase, the network will be unable to process new packets, and delays and traffic congestion may occur. The configuration of the communication terminal 2 will be described in detail later.

[0020] The communication control system according to this embodiment employs a mixed probability model in which the usage status of wireless resources is modeled using a mixed Bernoulli distribution. The communication control system uses the mixed probability model, in which parameters are set to reflect the intended communication traffic state, such as traffic leveling, as true data for adversarial learning, and trains a generator 121 that generates pseudo data similar to the true data. The communication control system also notifies communication terminal 2 of the trained generator 121' as communication control information indicating whether each time slot (resource block) is available for use. Each communication terminal 2 then transmits a signal using an available time slot.

[0021] [Function block of communication control device] Next, functional blocks of the communication control device 1 according to this embodiment will be described with reference to the block diagram of Fig. 1. As shown in Fig. 1, the communication control device 1 includes a first setting unit 10, a second setting unit 11, a learning unit 12, a first storage unit 13, and a notification unit 14.

[0022] The first setting unit 10 sets a distribution of observed data, which is a set of observed values, for a mixed probability model in which the usage status of each time slot into which wireless resources are divided is set as the observed value of each cluster. In this embodiment, a model following a mixed Bernoulli distribution is adopted as the mixed probability model. The observed value is a value that represents, as binary data, whether or not a signal is transmitted by communication terminal 2 for each time slot.

[0023] The mixed Bernoulli distribution employed in this embodiment is a model for clustering a set of N pieces of observed data formed by a vector of length D whose elements are 0 and 1. In this embodiment, the usage status of wireless resources is considered as a plurality of clusters, and the observed data is a set of observed values ​​that represent the presence or absence of signal transmission in each of a plurality of time slots with 0 and 1. The observed values ​​are generated from a plurality of different clusters having a binary distribution.

[0024] Here, we have N binary vector observations x1,x2,…,x NThe set of observed data X is given by the parameters μ1,μ2,…,μ K The set of observed data X follows a mixed Bernoulli distribution with a parameter data set M. That is, it is assumed that the observed data X is generated from multiple clusters, and each cluster is associated with a parameter μ that represents the average probability (average probability) of a signal being emitted.

[0025] The observation data X and the parameter data set M are expressed as the following equations (1) and (2), respectively (T is a transpose symbol).

number

[0026] nth binary vector observation x n and the kth parameter vector μ k Each of these is further expressed as a vector having D elements, as shown in the following equations (3) and (4).

number

[0027] Observation x n element x of n As described above, each of (i) (i=1, 2, ..., D) is a binary variable that takes the value of 0 or 1. In this embodiment, the element x n (i) is defined as follows:

number

[0028] The probability of transmitting a signal and the probability of not transmitting a signal are expressed by the following equations (6) and (7), respectively.

number

[0029] From the above equations (6) and (7), the observed value x of the observed data X nWhen cluster k is assigned to observation x n is the parameter μ of cluster k k It follows the Bernoulli distribution of the following equation (8):

number

[0030] Here, the parameter data set π of the occurrence frequency of each cluster (K ​​clusters) is defined as follows: The parameter π is the observed value x n is a parameter that indicates how much of the data is distributed to each cluster. k is the observed value x of the observed data X n is the mixture ratio occurring in cluster k.

number

[0031] Also, the observed value x n When cluster k is assigned to observation x n is μ k It follows D Bernoulli distributions with parameters. Furthermore, each observation x n are independently generated by the mixed Bernoulli distribution of the following equation (10).

number

[0032] The parameter μ, which indicates the probability of the Bernoulli distribution in each cluster, is defined by the definition of the mixed Bernoulli distribution in equation (10). k and the cluster mixing ratio π k and from all clusters, observation x n The probability of occurrence of the above equation (10) is a mixed probability model in this embodiment, and represents the usage status of 1 to D time slots of the radio resource.

[0033] The first setting unit 10 sets 1 to D observation values ​​xn The number and frequency of elements that take on the value "1" indicating the transmission of a signal can be set according to a desired usage pattern of the first to D time slots. Also, the first setting unit 10 sets the number and frequency of the observed values ​​x that take on the value "1" indicating the transmission of a signal across the plurality of clusters 1 to K. n Therefore, the first setting unit 10 can set the distribution of the observation data X that reflects the desired usage status of the wireless resource for each of a plurality of different clusters. n By setting the occurrence probability of , it becomes possible to set whether or not each communication terminal 2 is allowed to transmit a signal in 1 to D time slots.

[0034] FIG. 2 is a diagram for explaining the setting of the distribution of observation data X by the first setting unit 10. Clusters 1 to K shown in FIG. 2 represent the entire available radio resources. "1 to D" constituting each cluster represent the observed values ​​x indicating whether or not a signal is transmitted using 1 to D time slots. n As shown in Figure 2, for each data 1 to D in each cluster, time slots where signal transmission is permitted ("transmission" indicated by black dots in the figure) and time slots where signal transmission is not permitted (data other than the black dots in the figure) are set.

[0035] In the example of FIG. 2, one of time slots 1 to D is assigned to each communication terminal 2 in each of multiple clusters 1 to K. The times of time slots 1 to D in each cluster correspond to one another, and are set so that the times of time slots used by each communication terminal 2 when transmitting a signal do not overlap between multiple clusters 1 to K. In this way, the first setting unit 10 can set clusters that transmit signals with equal probability in each of the 1 to D time slots. The first setting unit 10 can set a specific value of the observation data X that equalizes communication traffic, as well as a value of the observation data X that reflects a specific usage pattern of the time slots in accordance with the specific design of the communication network, etc.

[0036] Returning to FIG. 1, the second setting unit 11 sets the parameters of the mixed probability model so as to obtain the distribution of the observed data X set by the first setting unit 10. More specifically, the second setting unit 11 calculates the parameter μ k , π k Using the example of FIG. 2 described above, when the first setting unit 10 sets clusters that transmit signals with equal probability in 1 to D time slots, the second setting unit 11 calculates the value of the parameter μ k , π k In this way, we can extract the observation x from each cluster. n The probability of occurrence of observation x from each cluster can be made equal. n The mixed probability model in which the probability of occurrence of is set in advance is used as true data during adversarial learning by the learning unit 12 described below.

[0037] The learning unit 12 learns the parameter μ set by the second setting unit 11. k , π k The adversarial learning of a generative model having a generator 121 that generates pseudo-usage information similar to the true usage information as true usage information indicating the time slots to be allocated to the communication terminal 2, and a classifier 122 that distinguishes between the pseudo-usage information generated by the generator 121 and the true usage information is performed using a mixed probability model having the above formula:

[0038] As shown in FIG. 3, the learning unit 12 adversarially trains a GAN (Generative Adversarial Network) having a generator 121 and a classifier 122. In this embodiment, it is possible to provide as many pairs of generators 121 and classifiers 122 as there are clusters (K pairs). Alternatively, as shown in FIG. 3, it is possible to provide one classifier 122 for K generators 121_1, . . . , 121_K. Through learning by the learning unit 12, K trained generators 121′ (trained generators 121′_1, . . . , 121′_K) are constructed.

[0039] 4 and 5 are diagrams schematically illustrating the neural network configuration of the generator 121 and the classifier 122 of the GAN used by the learning unit 12. As shown in FIG. 4, the generator 121 is configured as a neural network having an input layer, a hidden layer, and an output layer. Note that the K generators 121_1, . . . , 121_K each have the same network structure, and will be collectively referred to as the generator 121 below. The generator 121 is a model that generates pseudo-usage information from random noise. For example, m randomly sampled Gaussian noise vectors (z1 to z m ).

[0040] The generator 121 performs a product-sum operation on the input and weight parameters and performs threshold processing using an activation function to output an output G(z). The output G(z) from the generator 121 is calculated based on the parameter μ set by the second setting unit 11. k , π k The data is similar to the observed data X obtained by a mixture probability model having the set parameter μ k , π k The 1-D observations x of each cluster obtained by a mixture probability model with n Each output G(z1) to G(z D ) is the number of 1 to D observations x in one cluster. n As the neural network that configures the generator 121, a CNN or a ResNet can be used.

[0041] 5 is configured as a neural network having an input layer, a hidden layer, and an output layer. In the example of FIG. 5, the parameter μ k , π k The 1-D observations x in each cluster are obtained by a mixture probability model with n is given.

[0042] The classifier 122 performs a product-sum operation on the input and weight parameters and threshold processing using an activation function, and outputs a binary output of 1 or 0. When the classifier 122 correctly identifies the training data related to the input true usage information as true usage information, it outputs an output y=1. On the other hand, when the classifier 122 correctly identifies the training data related to the input pseudo usage information as pseudo usage information, it outputs an output y=0. In this way, the classifier 122 is a model that distinguishes the model distribution generated by the generator 121 from the data distribution of the training data, which is the true distribution. A CNN can be used as the neural network that constitutes the classifier 122.

[0043] FIG. 3 is a block diagram for explaining adversarial learning of GAN by the learning unit 12. The generator 121 of the GAN adopted by the learning unit 12 is represented by function G, and the discriminator 122 is represented by function D. Furthermore, the true usage information is represented by x, the predicted value output by the discriminator 122 is represented by y, and the correct label is represented by t. The correct label t is set to 1 for the true usage information and 0 for the pseudo usage information generated by the generator 121. At this time, the discriminator 122 calculates the cross entropy E CE It can be expressed as:

[0044]

number

[0045] The first term in the brace of the above equation (11) represents t n lny n In this case, the predicted value y nis the correct label of the true usage information, t n = 1. On the other hand, the second term in the braces represents (1-t n )ln(1-y n ), the predicted value y n is the correct label value (1-t n ) = 0. In this way, the cross entropy E CE is the maximum value when the predicted value matches the correct label value.

[0046] Here, the generator 121 that constitutes the GAN has parameters w G ,θ G and the function G(w G ,θ G ) The classifier 122 also uses the parameter w D ,θ D and function D(w D ,θ D ) The cross entropy E in the above equation (11) CE The objective function E of the GAN including the generator 121 and the discriminator 122 based on the above can be expressed by the following equation (12).

number

[0047] The first term of the above equation (12) represents E D(x)=1 lnD(w D ,θ D ) is the expected value that the classifier 122 will classify the true usage information as the true usage information. D(x)=0 ln(1-D(G(w G ,θ G ),w D ,θ D)) is the expected value at which the discriminator 122 discriminates the pseudo usage information generated by the generator 121 as pseudo usage information. In GAN learning, the generator 121 and the discriminator 122 are trained adversarially through min-max optimization of the objective function E. Therefore, the generator 121 is trained to be able to generate pseudo usage information that deceives the discriminator 122, and the discriminator 122 is trained to discriminate the pseudo usage information generated by the generator 121 as pseudo usage information.

[0048] In learning of the classifier 122, when true usage information is given, the classifier 122 outputs an output close to y=1, thereby maximizing the first term of the objective function E in the above equation (12). On the other hand, when pseudo usage information is given, the classifier 122 learns to output an output close to y=0, thereby maximizing the second term of the objective function E.

[0049] In the learning of the generator 121, D(G(w G ,θ G ),w D ,θ D ) (D(G(z)) in Figure 3) is close to 1. G ,θ G ) (G(z) in FIG. 3 ), thereby minimizing the objective function E. The learning unit 12 uses a learning procedure that alternately updates the parameters of the generator 121 and the classifier 122. Details of the learning procedure of the generator 121 and the classifier 122 by the learning unit 12 will be described later.

[0050] The first storage unit 13 stores a trained generator 121′ in which the objective function E of the GAN has been optimized by the learning unit 12. More specifically, the first storage unit 13 stores K trained generators 121′ (trained generators 121′_1, . . . , 121′_K) corresponding to clusters 1 to K. The first storage unit 13 also stores a parameter μ k , π k The mixed probability model of the above equation (10) is stored.

[0051] The notification unit 14 notifies the communication terminal 2 of the trained generator 121′ constructed by the learning unit 12 as communication control information. More specifically, the notification unit 14 can notify, for example, each of K communication terminals 2 via the network NW of any one trained generator 121′ among 1 to K trained generators 121′_1, . . . , 121′_K as communication control information.

[0052] [Communication terminal function block] Next, a description will be given of the functional blocks of the communication terminal 2. As shown in Fig. 1, the communication terminal 2 includes an acquisition unit 20, a second storage unit 21, a generation unit 22, an identification unit 23, and a communication control unit 24.

[0053] The acquisition unit 20 acquires the communication control information notified from the communication control device 1. Specifically, the acquisition unit 20 can acquire the learned generator 121′ of the communication control information stored in the second storage unit 21.

[0054] The second storage unit 21 stores communication control information notified from the communication control device 1. Specifically, the second storage unit 21 stores a learned generator 121′. The learned generator 121′ is one of K learned generators 121′_1, . . . , 121′_K constructed by learning by the learning unit 12 of the communication control device 1.

[0055] The generating unit 22 generates pseudo-usage information using the trained generator 121′ based on the communication control information notified from the communication control device 1. The pseudo-usage information generated by the generating unit 22 is the observed value x n As shown in the data set, the pseudo-usage information is information in which the value of "1" for transmitting a signal and the value of "0" for not transmitting a signal are assigned to each of the 1st to Dth time slots. For example, in the example of "cluster 1" in FIG. 2, the pseudo-usage information allows the transmission of a signal in the 1st slot, and does not allow the transmission of signals in the 2nd to Dth slots (x n =[x n (1), x n (2), x n (3),···,xn (D)]=[1,0,0,···,0]).

[0056] The specifying unit 23 specifies the time slots used by the communication terminal 2 based on the pseudo-use information generated by the generating unit 22. Specifically, the specifying unit 23 specifies the time slots used by the communication terminal 2 based on the pseudo-use information generated by the generating unit 22. n Among (i), a time slot having a value of 1 is identified. The time slot identified by the identification unit 23 is a time slot that is permitted to be used when the communication terminal 2 transmits a signal.

[0057] The communication control unit 24 issues an instruction to transmit a signal using the identified time slot. Specifically, the communication control unit 24 can instruct the processor 202 of the communication terminal 2 to synchronize timing with a base station (not shown), generate a signal, and transmit the signal. In response to the communication control instruction from the communication control unit 24, the communication terminal 2 transmits a signal using the time slot that is permitted to be used.

[0058] [Hardware configuration of communication control device] Next, an example of a hardware configuration for realizing the communication control device 1 having the above-described functions will be described with reference to FIG.

[0059] 6, the communication control device 1 can be realized by, for example, a computer including a processor 102, a main memory device 103, a communication interface 104, an auxiliary memory device 105, and an input / output (I / O) 106 connected via a bus 101, and a program that controls these hardware resources. The communication control device 1 further includes a display device 107.

[0060] The processor 102 is realized by a CPU, a GPU, an FPGA, an ASIC, or the like.

[0061] The main memory device 103 pre-stores programs for the processor 102 to perform various controls and calculations. The processor 102 and the main memory device 103 implement the functions of the communication control device 1, such as the first setting unit 10, the second setting unit 11, the learning unit 12, and the notification unit 14 shown in FIG.

[0062] The communication interface 104 is an interface circuit for connecting the communication control device 1 to various external electronic devices via a network.

[0063] The auxiliary storage device 105 is composed of a readable / writable storage medium and a drive for reading and writing various information such as programs and data from and to the storage medium. The auxiliary storage device 105 can use a semiconductor memory such as a hard disk or flash memory as the storage medium.

[0064] The auxiliary storage device 105 has a program storage area for storing a communication control program that generates communication control information and notifies the communication terminal 2. The auxiliary storage device 105 also has a program storage area for storing a learning program for performing GAN adversarial learning executed by the communication control device 1. The auxiliary storage device 105 also has a program storage area for storing a mixed probability model calculation program that calculates a mixed probability model related to a mixed Bernoulli distribution. The auxiliary storage device 105 realizes the first storage unit 13 described in FIG. 1. Furthermore, for example, the auxiliary storage device 105 may have a backup area for backing up the above-mentioned data, programs, etc.

[0065] The input / output I / O 106 is an input / output device that inputs signals from external devices and outputs signals to external devices.

[0066] The display device 107 is configured by an organic EL display, a liquid crystal display, etc. The display device 107 can display true usage information on the screen.

[0067] [Hardware configuration of communication terminal] Next, an example of a hardware configuration for realizing the communication terminal 2 having the above-described functions will be described with reference to FIG.

[0068] 7, similar to the communication control device 1, the communication terminal 2 can be realized by a computer including a processor 202, a main memory device 203, a communication interface 204, an auxiliary memory device 205, and an input / output (I / O) 206 connected via a bus 201, and a program for controlling these hardware resources. Furthermore, the communication terminal 2 includes a display device 207.

[0069] The main memory device 203 pre-stores programs for the processor 202 to perform various controls and calculations. The processor 202, which is realized by a CPU, GPU, FPGA, ASIC, etc., and the main memory device 203 realize the functions of the communication terminal 2, such as the acquisition unit 20, generation unit 22, identification unit 23, and communication control unit 24 shown in Fig. 1. The main memory device 203 also stores IP addresses.

[0070] The communication interface 204 is an interface circuit for connecting the communication terminal 2 to various external electronic devices via a network.

[0071] The auxiliary storage device 205 is composed of a readable / writable storage medium and a drive for reading and writing various information such as programs and data from and to the storage medium. The auxiliary storage device 205 can use a hard disk or semiconductor memory such as a flash memory as the storage medium.

[0072] The auxiliary storage device 205 has a program storage area for storing a generation program that performs calculations on the trained generator 121' based on the communication control information. The auxiliary storage device 105 also has a program storage area for storing a communication control program for transmitting a signal in a specified time slot. The auxiliary storage device 205 implements the second storage unit 21 described in FIG. 1. Furthermore, the auxiliary storage device 205 may have, for example, a backup area for backing up the above-mentioned data, programs, etc.

[0073] The input / output I / O 206 is an input / output device that inputs signals from external devices and outputs signals to external devices.

[0074] The display device 207 is configured by an organic EL display, a liquid crystal display, or the like.

[0075] The SIM 208 is composed of an IC chip, a security module, etc., and includes identifier information such as the subscriber identification number (IMSI: International Mobile Subscriber Identity) of the user of the communication terminal 2, the telephone number of the subscriber user (MSISDN: Mobile Subscriber International Subscriber Directory Number), and the SIM card number (ICCID: Integrated Circuit Card Identifier). The communication terminal 2 can be uniquely identified by the IMSI of the SIM 208.

[0076] [Operation sequence of communication control system] Next, the operation of the communication control system including the communication control device 1 and communication terminal 2 having the above-described configuration will be described with reference to the sequence of FIG.

[0077] 8, first, the first setting unit 10 of the communication control device 1 sets the distribution of the observation data X desired in the mixed probability model (step S1). Specifically, the first setting unit 10 sets the observed value x of the observation data X, which specifies the number and frequency of time slots that are permitted to transmit signals among 1 to D time slots for each cluster, so that the intended usage status of wireless resources, such as a state where communication traffic is leveled, is reflected. n (x n =[x n (1), x n (2), x n (3),···,x n In this way, in step S1, the observed value xn The values ​​of the observed data X can be set so that the probability of occurrence is equal or has a specific bias.

[0078] Next, the second setting unit 11 adjusts the parameter μ of the mixture probability model so as to obtain the distribution of the observation data X set in step S1. k , π k Specifically, the second setting unit 11 sets the observed value x for each cluster in the mixed probability model of the above formula (10). n When the observed data X is given with the probability of occurrence of k , π k The parameter μ is adjusted and determined. k , π k The mixed probability model in which the above is set is stored in the first storage unit 13.

[0079] Next, the learning unit 12 performs a learning process (step S3). Specifically, the learning unit 12 performs a learning process on the parameter μ k , π k As true usage information regarding the usage status of time slots, a mixed probability model having the above formula is used, and adversarial learning of a GAN having a generator 121 that generates pseudo usage information similar to the true usage information, and a classifier 122 that distinguishes between the pseudo usage information generated by the generator 121 and the true usage information is performed. Through the learning process, K trained generators 121'_1, . . . , 121'_K corresponding to clusters 1 to K, respectively, are constructed and stored in the first storage unit 13. Details of the learning process in step S3 will be described later.

[0080] Next, the notification unit 14 notifies the communication terminal 2 of the trained generator 121′ constructed by the learning unit 12 in step S3 as communication control information via the network NW (step S4). The notification unit 14 can arbitrarily select one of the K trained generators 121_1, . . . , 121′_K constructed for each cluster, and notify each of the multiple communication terminals 2 of the selected one.

[0081] Next, when the communication terminal 2 is notified of the communication control information from the communication control device 1, the communication terminal 2 stores the learned generator 121′ of the communication control information in the second storage unit 21. After that, the acquisition unit 20 of the communication terminal 2 acquires the learned generator 121′ by reading it from the second storage unit 21 (step S5).

[0082] Next, the generator 22 generates pseudo-usage information using the trained generator 121' acquired in step S5 (step S6). After that, the identifier 23 identifies a time slot that the communication terminal 2 is permitted to use when transmitting a signal, based on the pseudo-usage information generated in step S6 (step S7). Next, the communication control unit 24 issues an instruction to transmit a signal using the time slot identified in step S7 (step S8).

[0083] In response to the instruction in step S8, processor 202 of communication terminal 2 transmits a signal using the identified time slot (step S9). Note that other communication terminals 2 similarly execute the processes from step S5 to step S9 and transmit signals using the identified time slots. In this way, by allocating identified time slots to communication terminal 2 in accordance with the communication control information, communication traffic due to signals transmitted by multiple communication terminals 2 can be brought into an intended state.

[0084] For example, as shown in FIG. 2, a communication terminal 2 that generates pseudo usage information for cluster 1 using learned generator 121'_1 transmits a signal using the first time slot. Another communication terminal 2 that generates pseudo usage information for cluster 2 using learned generator 121'_2 transmits a signal using the second time slot. A Kth communication terminal 2 that generates pseudo usage information for cluster K using learned generator 121'_K transmits a signal using the Dth time slot. Therefore, multiple communication terminals 2 can transmit signals with equal probability in each time slot and with equal probability between clusters. As a result, communication traffic is leveled, leading to efficient use of wireless resources.

[0085] Next, referring to FIG. 9, the learning process (step S3) by the learning unit 12 of the communication control device 1 described in FIG. 8 will be described. In the example of the learning process shown in FIG. 9, the learning unit 12 repeatedly learns the true usage information for each of the 1 to K clusters. First, the learning unit 12 inputs the true usage information as training data 124 into the discriminator 122, and the parameters w D , θ D of the discriminator 122 are learned and updated (step S20). As shown in the block diagram of the learning unit 12 in FIG. 3, the parameters μ k , π k of the mixture probability model having are used as the training data 124 input when learning the discriminator 122 as the true usage information regarding the usage status of the time slot. In step S20, first, as the true usage information, among clusters 1 to K, 1 to D observed values x n of cluster 1 are used as the training data 124.

[0086] In step S20, the learning unit 12 can cause the discriminator 122 to learn the true usage information using, for example, the error backpropagation method or the like.In step S20, a discriminator 122 that can discriminate the true usage information of cluster 1 as the true usage information is constructed in advance.

[0087] Next, the learning unit 12 generates Gaussian noise and gives a random vector of the generated Gaussian noise as an input to the generator 121 (step S21). Subsequently, the generator 121 performs a sum-of-products operation of the input z and the weight parameters w G , θ G and threshold processing by an activation function to generate pseudo usage information G(z) (step S22). In step S22, as shown in FIG. 3, the learning unit 12 synthesizes the 1 to K outputs generated by each of the K generators 121_1, ···, 121_K (point a shown in FIG. 3) to generate one pseudo usage information G(z).

[0088] Next, the learning unit 12 performs learning of the classifier 122. The learning of the classifier 122 is performed by using the parameter w D ,θ D First, the learning unit 12 performs the learning process by fixing the parameter μ set in step S2 of FIG. k , π k The true usage information is input to the classifier 122 as training data 124. Then, the learning unit 12 adjusts the parameters w by backpropagation or the like so that the objective function E in the above equation (12) is maximized. D ,θ D (Step S23). The label of the training data 124 is set to 1 (true usage information). In step S23, first, the learning unit 12 updates the 1 to D observation values ​​x n is used as the true utilization information to train the classifier 122.

[0089] Next, the learning unit 12 provides the pseudo-usage information generated by the generator 121 in step S22 to the discriminator 122 as an input, and calculates the parameter w by backpropagation or the like so that the objective function E in the above equation (12) is maximized. D ,θ D (Step S24). That is, in steps S23 and S24, in order to maximize the objective function E in the above equation (12), the first term is updated as D(w D ,θ D )=1 is output, and the second term is D(G(w G ,θ G ),w D ,θ D )=0. Note that the label 0 (pseudo usage information) is set in the training data 124. In step S24, the classifier 122 is trained using one pseudo usage information G(z) generated by each of the K generators 121_1, . . . , 121_K and combined.

[0090] The learning of the classifier 122 in steps S23 and S24 corresponds to the dashed arrows in the block diagram of the learning unit 12 shown in FIG. 3 , which indicate that a classifier error is calculated in block 125 of the objective function E based on output 123 from the classifier 122, and then the error is back-propagated to the classifier 122.

[0091] Next, the learning unit 12 trains the generator 121. The training of the generator 121 is performed with the parameters of the discriminator 122 fixed. The learning unit 12 trains the generator 121 so that pseudo-usage information is generated when random Gaussian noise is given to the generator 121. Specifically, the learning unit 12 trains the parameter w by backpropagation or the like in order to minimize the objective function E in the above equation (12). G ,θ G is updated (step S25).

[0092] The learning in step S25 corresponds to the flow of the dashed arrow indicating backpropagation of error to the generator 121 in the block diagram of the learning unit 12 in Fig. 3. That is, step S25 corresponds to the flow of the dashed arrow in which pseudo-usage information generated by the generator 121 in Fig. 3 is input to the discriminator 122, a generator error is calculated from the output 123 in the block 125 of the objective function E, and the error is further backpropagated to the generator 121.

[0093] Thereafter, learning of the discriminator 122 and the generator 121 (generators 121_1 to 121_K) from step S22 to step S25 is repeated until the value of the objective function E reaches a Nash equilibrium and converges (step S26: NO). On the other hand, if the value of the objective function E converges (step S26: YES), the processing from step S20 to step S26 is repeated using the remaining true usage information from cluster 2 to cluster K, out of all the true usage information from cluster 1 to cluster K, until learning of K generators 121_1, . . . , 121_K and the discriminator 122 is performed (step S27: NO).

[0094] Thereafter, when the generator 121 (generators 121_1, . . . , 121_K) and the discriminator 122 are trained using the true usage information of the remaining K−1 clusters from cluster 2 to cluster K (step S27: YES), the training unit 12 stores 1 to K trained generators 121′_1, . . . , 121′_K in the first storage unit 13 (step S28). The trained generator 121′ is constructed by the above-described processing from step S20 to step S28. After that, the processing proceeds to step S4 in FIG. 8.

[0095] As described above, according to the communication control device 1 of this embodiment, the usage status of wireless resources is modeled by a mixed Bernoulli distribution, and the parameter μ is adjusted so as to obtain a distribution of the observation data X that represents the intended communication traffic state. k , π k The mixed probability model with the above set is used as the true data in the adversarial learning of the GAN. Furthermore, the trained generator 121' constructed by the adversarial learning is notified to the communication terminal 2 as communication control information, so that communication control regarding the use of wireless resources can be performed on the IoT terminal so that the intended communication traffic state is achieved.

[0096] Furthermore, according to the communication control device 1 of this embodiment, the parameter μ k , π k The mixed probability model in which the above is set is used as true usage information, which is true data in the adversarial learning of the GAN, and a generator 121 that generates pseudo usage information similar to the true usage information is trained. Therefore, by performing the adversarial learning of the GAN, it is possible to learn the latent variables of the mixed probability model that follows the mixed Bernoulli distribution.

[0097] Furthermore, according to the communication control device 1 of this embodiment, the usage status of 1 to D time slots of one radio resource is set as the observed value of each cluster of a mixed Bernoulli distribution, so that it is possible to specify the usage pattern and frequency of the radio resource for each cluster. This makes it possible to set the usage pattern of the radio resource in more detail and realize more efficient resource management.

[0098] Furthermore, according to the communication control device 1 according to the present embodiment, the trained generator 121′ constructed by the adversarial learning of the GAN is notified as communication control information to the communication terminal 2 via the network NW. Therefore, communication traffic can be controlled remotely.

[0099] In the embodiment described above, the resource blocks of the radio resources are time slots, but the resource blocks may be different channels or spatial resources in a frequency band.

[0100] In the embodiment described above, the case where the communication control device 1 notifies the communication terminal 2 of the communication control information via the network NW has been exemplified. However, the communication control device 1 may be configured to notify each communication terminal 2 of the communication control information via a core network or a base station.

[0101] The above describes embodiments of the communication control device, communication terminal, and communication control method of the present invention, but the present invention is not limited to the described embodiments, and various modifications that a person skilled in the art can imagine are possible within the scope of the invention described in the claims. [Explanation of symbols]

[0102] 1...communication control device, 2...communication terminal, 10...first setting unit, 11...second setting unit, 12...learning unit, 13...first memory unit, 14...notification unit, 20...acquisition unit, 21...second memory unit, 22...generation unit, 23...identification unit, 24...communication control unit, 101, 201...bus, 102, 202...processor, 103, 203...main memory unit, 104, 204...communication interface, 105, 205...auxiliary memory unit, 106, 206...input / output I / O, 107, 207...display device, 121...generator, 122...identifier, 123...output, 124...training data, 125...block of objective function E, NW...network.

Claims

1. a first setting unit configured to set a distribution of observation data, which is a set of observation values, for a mixed probability model in which a utilization status of each resource block obtained by dividing a wireless resource is an observation value of each cluster; a second setting unit configured to set parameters of the mixed probability model so as to obtain a set distribution of the observation data; a learning unit configured to perform adversarial learning of a generative model including a generator that generates pseudo-usage information similar to true usage information regarding the usage status of the resource block using the mixed probability model having the set parameters as true usage information, and a classifier that distinguishes between the pseudo-usage information generated by the generator and the true usage information; a notification unit configured to notify a communication terminal of the trained generator constructed by the learning unit as communication control information; A communication control device comprising:

2. 2. The communication control device according to claim 1, the mixed probability model is a model that follows a mixed Bernoulli distribution, The parameters of the mixture probability model include a mixture ratio representing the probability that each cluster generates the observed value, and an average probability representing the average probability that the resource block is used in each cluster. A communication control device characterized by:

3. 2. The communication control device according to claim 1, The first setting unit sets a probability of occurrence of the observed value for each cluster. A communication control device characterized by:

4. 2. The communication control device according to claim 1, the resource block is a time slot obtained by dividing the radio resource in the time domain into a plurality of time slots, The observation value is a value that represents, as binary data, whether or not the communication terminal transmits a signal for each time slot. A communication control device characterized by:

5. a generating unit configured to generate the pseudo-usage information using the trained generator based on the communication control information notified from the communication control device according to any one of claims 1 to 4; an identification unit configured to identify a resource block to be used based on the pseudo-utilization information generated by the generation unit; a communication control unit that issues an instruction to transmit a signal using the identified resource block; A communication terminal comprising:

6. a first setting step of setting a distribution of observed data, which is a set of observed values, for a mixed probability model in which the utilization status of each resource block obtained by dividing a wireless resource is set as an observed value for each cluster; a second setting step of setting parameters of the mixed probability model so as to obtain the set distribution of the observation data; a learning step of performing adversarial learning of a generative model having a generator that generates pseudo-usage information similar to true usage information regarding the usage status of the resource block using the mixture probability model having the set parameters as true usage information, and a classifier that distinguishes between the pseudo-usage information generated by the generator and the true usage information; a notification step of notifying a communication terminal of the trained generator constructed in the learning step as communication control information; A communication control method comprising:

7. 7. The communication control method according to claim 6, The first setting step sets a probability of occurrence of the observed value for each cluster. A communication control method comprising:

Images