Estimation device, wireless communication system including the same, and program for causing a computer to execute the same

Abstract

To provide an estimation device capable of preparing a schedule for using a reflective element by determining in advance whether a reflective element sufficient to actually secure a sufficient transmission rate is available.SOLUTION: Estimating means 43 estimates an usable reflective element from among N reflective elements using a learning device 44 on the basis of first time series data consisting of the busy state and the idle state, and estimates the achievable rate of the IRS reflective element in a propagation path by using the learning device 45 for the reflective element used among the N reflective elements on the basis of second time series data showing the achievable rate of N reflective elements when a transmitter is transmitting signals using MT antennas. The estimating means 43 creates a communication schedule using the prediction result of the availability of the N reflective elements and the maximum achievable rate of the reflective element being used among the N reflective elements.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to an estimation device, a wireless communication system including the same, and a program to be executed by a computer. [Background technology]

[0002] Next-generation communications standards call for the creation of a communications infrastructure that will enable access to all people, information, and things anywhere, both domestically and internationally, by utilizing high-frequency bands such as millimeter waves and terahertz waves, while enhancing the high speed, large capacity, low latency, and multiple connection capabilities that are characteristic of fifth-generation mobile communications systems (5G).

[0003] High-frequency radio waves cannot be expected to propagate beyond line-of-sight, resulting in weak spots. For this reason, there has been active research into technology that solves the problem of high-frequency radio waves by using an intelligent reflecting surface (IRS), which is made up of multiple passive reflecting elements whose reflection characteristics can be dynamically changed (Non-Patent Document 1). By appropriately changing the reflection characteristics of each element and creating a radio wave propagation path that bypasses obstacles, stable communication can be achieved even beyond line-of-sight.

[0004] IRS basically acts as an electromagnetic wave reflector and does not have high-frequency circuits, so it has the advantage of consuming significantly less power than conventional multi-hop communications for expanding coverage. [Prior art documents] [Non-patent literature]

[0005] [Non-Patent Document 1] Q. Wu and R. Zhang, “Towards Smart and Reconfigurable Environment: Intelligent Reflecting Surface Aided Wireless Network,” IEEE Commun. Mag., vol. 58, no. 1, pp. 106-112, 2020. Summary of the Invention [Problem to be solved by the invention]

[0006] However, in a wireless communication system using an IRS (an IRS-assisted system), the achievable transmission rate is not predicted, and the number of idle slots in which the IRS's reflecting elements are not in use is also not predicted. Therefore, it is difficult to determine in advance whether enough reflecting elements are available to ensure a sufficient transmission rate and to create a reflecting element usage schedule.

[0007] Therefore, according to an embodiment of the present invention, an estimation device is provided that can create a schedule for using reflecting elements by determining in advance whether enough reflecting elements are available to actually ensure a sufficient transmission rate.

[0008] Furthermore, according to an embodiment of the present invention, a wireless communication system is provided that includes an estimation device that can determine in advance whether enough reflecting elements are actually available to ensure a sufficient transmission rate and create a schedule for using reflecting elements.

[0009] Furthermore, according to an embodiment of the present invention, a program is provided that causes a computer to make a prior determination as to whether enough reflecting elements are available to actually ensure a sufficient transmission rate and to create a schedule for using the reflecting elements. [Means for solving the problem]

[0010] (Configuration 1) According to an embodiment of the present invention, the estimation deviceT (M T a transmitter having N (N is an integer of 2 or more) antennas, a first IRS which is an electromagnetic wave reflector having N (N is an integer of 2 or more) reflecting elements whose reflection characteristics can be dynamically changed, and M R (M R The present invention relates to an estimation device for estimating information useful for predicting a usage schedule, which is a schedule for using N reflecting elements, in a wireless communication system in which a transmitter and a receiver each having (an integer of 2 or more) antennas are non-linearly arranged, and the estimation device comprises a collection means and an estimation means. The collection means collects first time series data indicating, for each of the N reflecting elements, time series data consisting of a busy state in which the reflecting element is in use and an idle state in which the reflecting element is not in use, and a time series data indicating, for each of the N reflecting elements, ... transmitter and a receiver each having (an integer of 2 or more) antennas are non-linearly arranged. T and second time series data indicating achievable rates, which are transmission rates achievable by the N reflecting elements when transmitting a signal using the antennas. The estimation means executes a first estimation process for estimating, for each of the N reflecting elements, a usability prediction result, which is a result of predicting whether or not a reflecting element in the first IRS is usable, using a first learning device, based on the first time series data, and a second estimation process for estimating, for each of the N reflecting elements, the achievable rates of the N reflecting elements of the first IRS on the propagation path, for reflecting elements that are in use among the N reflecting elements, using a second learning device, based on the second time series data.

[0011] (Configuration 2) In configuration 1, the estimation means, in a first estimation process, inputs first time series data to a first learning device and estimates usable reflecting elements among the N reflecting elements by receiving from the first learning device an idle state corresponding to the maximum probability among probabilities that a first class classified according to the idle state of the first time series data matches a first learning stage class, which is a class classified according to the idle state of the first learning time series data acquired in the learning stage; and, in a second estimation process, inputs second time series data to a second learning device and estimates the achievable rate of a used reflecting element among the N reflecting elements by receiving from the second learning device an achievable rate corresponding to the maximum probability among probabilities that a second class classified according to the achievable rate of the second time series data matches a second learning stage class, which is a class classified according to the achievable rate of the second learning time series data acquired in the learning stage.

[0012] (Configuration 3) In configuration 1 or 2, the first learning device consists of N first machine learning devices, the second learning device consists of E (E is the number of reflection patterns of one or more reflecting elements when one or more of the N reflecting elements are used simultaneously) second machine learning devices, the first time series data consists of first idle state / busy state data to Nth idle state / busy state data, each of which arranges idle states and busy states in time series, and the second time series data consists of first achievable rate data to Eth achievable rate data, each of which arranges achievable rates in time series, associated with the E reflection patterns.

[0013] The estimation means is In a first estimation process, first idle state / busy state data is input to one first machine learning device, and a first process is performed to receive from the one first machine learning device an idle state corresponding to the maximum probability among the probabilities that a first class classified according to the idle state of the first idle state / busy state data matches a first learning stage class, which is a class classified according to the idle state of the first idle state / busy state data of the learning stage acquired in the learning stage, and N usability prediction results of N reflecting elements are estimated by performing the same process as the first process for the second idle state / busy state data to the Nth idle state / busy state data instead of the first idle state / busy state data in the first process and all of the (N-1) first machine learning devices; In the second estimation process, the first achievable rate data is input into one second machine learning machine, and a second process is performed in which an achievable rate corresponding to the maximum probability among the probabilities that the first class classified according to the achievable rate of the first achievable rate data matches a second learning stage class, which is a class classified according to the achievable rate of the first achievable rate data of the learning stage obtained in the learning stage, is received from the one first machine learning machine, and the achievable rate of the N reflecting elements used is estimated by performing the same process as the second process for the second achievable rate data to the Eth achievable rate data and all (E-1) second machine learning machines, instead of the first achievable rate data in the second process.

[0014] (Configuration 4) In configuration 3, N first machine learning devices execute the first process in parallel in the first estimation process, and E second machine learning devices execute the second process in parallel in the second estimation process.

[0015] (Configuration 5) In configuration 3 or 4, each of the N first machine learning devices classifies the idle states acquired in the first learning stage into a first plurality of classes and stores them in a first prediction matrix, and when one piece of idle state / busy state data is input, selects one class from the first plurality of classes stored in the first prediction matrix that has the highest probability of matching the idle state of the input one piece of idle state / busy state data, and outputs a usability prediction result to the estimation means based on the selection of one class; Each of the E second machine learning devices classifies the achievable rates obtained in the second learning stage into a second plurality of classes and stores them in a second prediction matrix. When a single achievable rate data is input, the device selects from the second plurality of classes one class that has the highest probability of matching the achievable rate of the input achievable rate data, and outputs the achievable rate classified into the selected class to the estimation means.

[0016] (Configuration 6) In any one of the configurations 1 to 5, the estimation means estimates M of the receiver in the second estimation process. R The propagation paths to the antennas are divided into M independent R Transformed into M independent propagation paths R M propagation paths R Calculate the sending rate of M R The sum of these transmission rates is the achievable rate.

[0017] (Configuration 7) In the sixth aspect, the estimation means, in the second estimation process, R Based on the signal-to-interference-plus-noise power ratio of the antennas, M R Calculate the transmission rate.

[0018] (Configuration 8) According to an embodiment of the present invention, a wireless communication system includes a first estimation device comprising the estimation device according to any one of configurations 1 to 7, a transmitter, a receiver, and a first IRS. The first estimation device is disposed within the first IRS.

[0019] (Configuration 9) In an eighth embodiment, the wireless communication system further includes a second estimator and a second IRS. The second estimator is the estimator according to any one of the first to seventh embodiments. The second IRS is an N-channel estimator with dynamically variable ray characteristics. * (N * is an electromagnetic wave reflector having reflective elements (where n is an integer greater than or equal to 2). The second estimator is disposed within the second IRS.

[0020] (Configuration 10) Further, according to an embodiment of the present invention, the program M T (M T a transmitter having N (N is an integer of 2 or more) antennas, a first IRS which is an electromagnetic wave reflector having N (N is an integer of 2 or more) reflecting elements whose reflection characteristics can be dynamically changed, and M R (M R A program for causing a computer to execute estimation of information useful for predicting a usage schedule, which is a schedule for using N reflecting elements, in a wireless communication system in which a receiver having N (an integer greater than or equal to 2) antennas is non-linearly arranged, the program comprising: The collecting means collects first time series data indicating time series data for each of the N reflecting elements, the time series data being composed of a busy state in which the reflecting element is in use and an idle state in which the reflecting element is not in use, and T a first step of collecting time series data indicating achievable rates for the N reflecting elements, the achievable transmission rates of the N reflecting elements when transmitting signals using the N antennas; The program causes a computer to execute a second step in which an estimation means executes a first estimation process in which a usability prediction result is estimated for each of N reflecting elements, the usability prediction result being the result of predicting whether or not a reflecting element in a first IRS is usable using a first learning device based on one time series data, and a second estimation process in which a second learning device is used to estimate the achievable rates of the N reflecting elements of the first IRS in the propagation path for those reflecting elements that are in use among the N reflecting elements, based on second time series data.

[0021] (Configuration 11) In configuration 10, the estimation means, in a first estimation process of the second step, inputs first time series data to a first learning device and estimates usable reflecting elements among the N reflecting elements by receiving from the first learning device an idle state corresponding to the maximum probability among probabilities that a first class classified according to the idle state of the first time series data matches a first learning stage class, which is a class classified according to the idle state of the first learning time series data obtained in the learning stage; and, in a second estimation process of the second step, inputs second time series data to a second learning device and estimates the achievable rate of a used reflecting element among the N reflecting elements by receiving from the second learning device an achievable rate corresponding to the maximum probability among probabilities that a second class classified according to the achievable rate of the second time series data matches a second learning stage class, which is a class classified according to the achievable rate of the second learning time series data obtained in the learning stage.

[0022] (Configuration 12) In configuration 10 or 11, the first learning device consists of N first machine learning devices, the second learning device consists of E (E is the number of reflection patterns of one or more reflecting elements when one or more of the N reflecting elements are used simultaneously) second machine learning devices, the first time series data consists of first idle state / busy state data to Nth idle state / busy state data, each of which arranges idle states and busy states in a time series, and the second time series data consists of first achievable rate data to Eth achievable rate data, each of which corresponds to E reflection patterns and arranges achievable rates in a time series.

[0023] The estimation means is In the first estimation process of the second step, the first idle state / busy state data is input to one first machine learning device, and a first process is performed to receive from the one first machine learning device an idle state corresponding to the maximum probability among the probabilities that a first class classified according to the idle state of the first idle state / busy state data matches a first learning stage class, which is a class classified according to the idle state of the first idle state / busy state data of the learning stage acquired in the learning stage, and the same process as the first process is performed for the second idle state / busy state data to the Nth idle state / busy state data instead of the first idle state / busy state data in the first process and all of the (N-1) first machine learning devices, thereby estimating N usability prediction results of the N reflecting elements; In the second estimation process of the second step, the first achievable rate data is input into one second machine learning machine, and a second process is performed to receive from the one first machine learning machine an achievable rate corresponding to the maximum probability among the probabilities that the first class classified according to the achievable rate of the first achievable rate data matches a second learning stage class, which is a class classified according to the achievable rate of the first achievable rate data of the learning stage obtained in the learning stage, and the achievable rate of the reflecting element being used among the N reflecting elements is estimated by performing the same process as the second process for the second achievable rate data to the Eth achievable rate data and all (E-1) second machine learning machines, instead of the first achievable rate data in the second process.

[0024] (Configuration 13) In configuration 12, N first machine learning machines execute the first process in parallel in the first estimation process of the second step, and E second machine learning machines execute the second process in parallel in the second estimation process of the second step.

[0025] (Configuration 14) In configuration 12 or 13, each of the N first machine learning devices classifies the idle states acquired in the first learning stage into a first plurality of classes and stores them in a first prediction matrix, and when one idle state / busy state data is input, selects one class from the first plurality of classes stored in the first prediction matrix that has the highest probability of matching the idle state of the input one idle state / busy state data, and outputs a usability prediction result to the estimation means based on the selection of one class; Each of the E second machine learning devices classifies the achievable rates obtained in the second learning stage into a second plurality of classes and stores them in a second prediction matrix. When a single achievable rate data is input, the device selects from the second plurality of classes one class that has the highest probability of matching the achievable rate of the input achievable rate data, and outputs the achievable rate classified into the selected class to the estimation means.

[0026] (Configuration 15) In any of the configurations 10 to 14, the estimation means estimates M of the receiver in the second estimation process of the second step. R The propagation paths to the antennas are divided into M independent R Transformed into M independent propagation paths R M propagation paths R Calculate the sending rate of M R The sum of these transmission rates is the achievable rate.

[0027] (Configuration 16) In the configuration 15, the estimation means calculates M R Based on the signal-to-interference-plus-noise power ratio of the antennas, M R Calculate the transmission rate. [Effects of the Invention]

[0028] A schedule for using the reflecting elements can be created by determining in advance whether or not enough reflecting elements are actually available to ensure a sufficient transmission rate. [Brief explanation of the drawings]

[0029] [Figure 1] 1 is a schematic diagram of a wireless communication system according to an embodiment of the present invention; [Figure 2] FIG. 2 is a schematic diagram of the IRS controller shown in FIG. 1. [Figure 3] FIG. 1 is a conceptual diagram of time series data TSD_1(t). [Figure 4] FIG. 1 is a conceptual diagram of time series data TSD_2(t). [Figure 5] FIG. 3 is a schematic diagram of a machine learning device that constitutes the learning device 44 shown in FIG. 2. [Figure 6] FIG. 6 is a schematic diagram of the probabilistic neural network PNNSLT_1 shown in FIG. 5. [Figure 7] FIG. 10 is a conceptual diagram showing a prediction matrix PSLT r that predicts usable reflective elements. [Figure 8] FIG. 8 is a schematic diagram showing a specific example of the prediction matrix PSLT r shown in FIG. 7. [Figure 9] FIG. 3 is a schematic diagram of a machine learning device that constitutes the learning device 45 shown in FIG. 2. [Figure 10] FIG. 10 is a conceptual diagram showing a prediction matrix PARr for predicting an achievable rate AR. [Figure 11] 3 is a flowchart illustrating the operation of the estimation device shown in FIG. 2. [Figure 12] 12 is a flowchart for explaining detailed operations of step S4 shown in FIG. 11. [Figure 13] 12 is a flowchart for explaining detailed operations of step S5 shown in FIG. 11. [Figure 14] FIG. 1 is a schematic diagram of another wireless communication system according to an embodiment of the invention. [Figure 15] 1 is a schematic diagram of learners 44 and 45 when an IRS controller 22 of an IRS 20 controls 2N reflecting elements 21. FIG. [Figure 16] 15 is a schematic diagram illustrating one specific example of the wireless communication system shown in FIG. 14. DETAILED DESCRIPTION OF THE INVENTION

[0030] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will now be described in detail with reference to the accompanying drawings, in which like or corresponding parts are designated by like reference numerals and will not be described repeatedly.

[0031] Fig. 1 is a schematic diagram of a wireless communication system according to an embodiment of the present invention. Referring to Fig. 1, a wireless communication system 10 according to the embodiment of the present invention includes a transmitter 1, an IRS 2, and a receiver 3. The IRS 2 includes N (N is an integer equal to or greater than 2) reflecting elements 21 and an IRS controller 22. The N reflecting elements 21 are arranged, for example, in a grid pattern. That is, the N reflecting elements 21 are arranged at equal intervals. The IRS controller 22 includes an estimation device 4.

[0032] Transmitter 1, IRS2, and receiver 3 are arranged in a wireless communication space. Transmitter 1, IRS2, and receiver 3 are arranged in positions where the arrangement [transmitter 1-IRS2-receiver 3] is not on a straight line. More specifically, when the line connecting transmitter 1 and IRS2 is L1 and the line connecting IRS2 and receiver 3 is L2, transmitter 1, IRS2, and receiver 3 are arranged so that the angle α formed by line L1 and line L2 is greater than 0. In other words, transmitter 1, IRS2, and receiver 3 are arranged non-linearly in the wireless communication space.

[0033] Transmitter 1 is M T (M T is an integer equal to or greater than 2), and the receiver 3 has M R (M R has antennas (an integer equal to or greater than 2). T is M R may be the same as M R may be different from.

[0034] Transmitter 1 is M T The IRS 2 transmits signals using N antennas. In the IRS 2, the IRS controller 22 controls the reflection coefficients of the N reflecting elements 21 to control the reflection angle and phase of the radio waves by the N reflecting elements 21. The N reflecting elements 21 then reflect the radio waves incident from the transmitter 1 toward the receiver 3. As a result, the radio waves reflected by the N reflecting elements 21 reach the receiver 3.

[0035] The receiver 3 receives the radio waves transmitted from the transmitter 1. In the wireless communication system 10, the propagation path between the transmitter 1 and the receiver 3 consists of a direct link where the radio waves from the transmitter 1 reach the receiver 3 directly, and an indirect link where the radio waves from the transmitter 1 reach the receiver 3 via the IRS 2.

[0036] When there is no obstacle OBT between the transmitter 1 and the receiver 3, the receiver 3 receives the radio waves transmitted from the transmitter 1 using a direct link and an indirect link.

[0037] On the other hand, if an obstacle OBT exists between the transmitter 1 and the receiver 3, the receiver 3 receives the radio waves transmitted from the transmitter 1 using an indirect link.

[0038] Therefore, in the wireless communication system 10, even if an obstacle OBT exists between the transmitter 1 and the receiver 3, radio waves can be transmitted from the transmitter 1 to the receiver 3 by passing through the IRS2.

[0039] The receiver 3 receives the signal transmitted from the transmitter 1, detects the signal to noise ratio (SNR), which is the ratio of the signal power to the noise power when the signal is received, and transmits the detected signal to noise ratio (SNR) to the IRS controller 22.

[0040] The estimation device 4 collects N pieces of time series data D1_chr_1(t) to D1_chr_N(t) showing time series data D1_chr(t) for N reflecting elements 21 of the IRS2, which consists of a busy state B in which the reflecting elements 21 are in use and an idle state I in which the reflecting elements 21 are not in use, as time series data TSD_1(t).

[0041] The estimation device 4 also detects whether the transmitter 1 is M T N pieces of time series data D2_chr_1(t) to D2_chr_N(t) showing the achievable rate, which is the achievable transmission rate of one reflecting element when transmitting a signal using N antennas, for N reflecting elements 21 are collected as time series data TSD_2(t).

[0042] Then, the estimation device 4 estimates usability prediction results for the N reflecting elements 21, which are the results of predicting whether each of the N reflecting elements 21 in the IRS2 is usable or not, based on the N time series data D1_chr_1(t) to D1_chr_N(t) using a method described below.

[0043] In addition, the estimation device 4 estimates the achievable rate of one reflecting element 21 of the IRS2 for N reflecting elements 21 on the propagation path from the transmitter 1 to the receiver 3 via the IRS2 based on N pieces of time series data D2_chr_1(t) to D2_chr_N(t) using a method described below.

[0044] Furthermore, the estimation device 4 estimates the frequency response of the propagation path between the transmitter 1 and the receiver 3 by a method described later. In this case, the transmitter 1 transmits a pilot signal, which is a known signal, and the estimation device 4 receives a received signal y(t) from the receiver 3 when the receiver 3 receives the pilot signal via the IRS controller 22. Then, the estimation device 4 estimates the frequency response of the propagation path based on the received received signal y(t).

[0045] Fig. 2 is a schematic diagram of the IRS controller 22 shown in Fig. 1. Referring to Fig. 2, the IRS controller 22 includes an antenna 221, a communication means 222, a control means 223, and an estimation device 4. The estimation device 4 includes a collection means 41, a storage means 42, an estimation means 43, and learners 44 and 45.

[0046] The communication means 222 of the IRS controller 22 receives the received signal y(t) of the pilot signal from the receiver 3 via the antenna 221, and outputs the received signal y(t) of the pilot signal to the estimation means 43 of the estimation device 4.

[0047] In addition, the communication means 222 of the IRS controller 22 receives, via the antenna 221, the signal-to-noise ratio SNR when the receiver 3 receives the received signal y(t) of the pilot signal from the receiver 3, and outputs the received signal-to-noise ratio SNR to the collection means 41 of the estimation device 4.

[0048] Furthermore, when the communication means 222 of the IRS controller 22 receives an instruction signal INS_1 from the control means 223 to instruct the transmitter 1 to transmit a training signal, the communication means 222 transmits the instruction signal INS_1 to the transmitter 1 via the antenna 221 .

[0049] When a signal is transmitted from the transmitter 1 to the receiver 3 via the IRS 2, the control means 223 of the IRS controller 22 detects, for each of the N reflecting elements, a period ON_period during which the reflecting element 21 is turned on and a period OFF_period during which the reflecting element 21 is turned off in time series. Then, the control means 223 executes, for the N reflecting elements 21, generating a period ON_period / period OFF_period(t) in which the period ON_period and the period OFF_period are arranged in time series, to generate N periods ON_period / period OFF_period_1(t) to period ON_period / period OFF_period_N(t), and outputs the generated N periods ON_period / period OFF_period_1(t) to period ON_period / period OFF_period_N(t) to the collection means 41 of the estimation device 4.

[0050] Furthermore, the control means 223 of the IRS controller 22 calculates the maximum achievable rate R among the achievable rates calculated by applying each reflection beamforming vector in the codebook to each combination of available reflection elements (combinations possible among reflection elements predicted to be in the OFF state) among the combinations of the N reflection elements 21 of the IRS 2. max and the maximum achievable rate R max The combination of reflecting elements that gives the reflected beamforming vector φ max and the control means 223 of the IRS controller 22 receives the maximum achievable rate R max and the maximum achievable rate R max The combination of reflecting elements that gives the reflected beamforming vector φ max and are aggregated.

[0051] The control means 223 of the IRS controller 22 determines the maximum achievable rate R max and the maximum achievable rate R max The combination of reflecting elements that gives the reflected beamforming vector φ max Aggregating these gives the maximum achievable rate R max and the maximum achievable rate Rmax The combination of reflecting elements that gives the reflected beamforming vector φ max and the maximum achievable rate R max / maximum achievable rate R max The combination of reflecting elements / reflecting beamforming vector φ max ] is held.

[0052] The control means 223 of the IRS controller 22 then determines the maximum achievable rate R max When transmitting a signal at , the reflected beamforming vector φ max By controlling the reflection coefficients of the N reflecting elements 21 based on the above, the reflection angle and phase of the radio wave by the N reflecting elements 21 are controlled.

[0053] When the collection means 41 receives N periods ON_period / period OFF_period_1(t) to period ON_period / period OFF_period_N(t) from the control means 223 of the IRS controller 22, it performs the generation of time series data for the N periods ON_period / period OFF_period_1(t) to period ON_period / period OFF_period_N(t) in which the period ON_period is in busy state B and the period OFF_period is in idle state I, thereby generating time series data TSD_1(t), and stores the generated time series data TSD_1(t) in the storage means 42.

[0054] Furthermore, the collecting means 41 receives the signal-to-noise ratio SNR from the communication means 222 of the IRS controller 22, and receives the frequency response FQR_0 of the propagation path between the transmitter 1 and the receiver 3 from the estimation means 43. Then, every time the collecting means 41 receives the signal-to-noise ratio SNR, it calculates the achievable rate AR based on the signal-to-noise ratio SNR and the frequency response FQR_0 by a method described later, generates time series data TSD_2(t) in which the achievable rate AR is associated with the calculation time point t of the achievable rate AR, and stores the generated time series data TSD_2(t) in the storage means 42.

[0055] The storage means 42 stores the time series data TSD_1(t) and the time series data TSD_2(t).

[0056] The estimation means 43 receives the received signal y(t) from the communication means 222 of the IRS controller 22, estimates the frequency response FQR_0 based on the received signal y(t) using a method described later, holds the estimated frequency response FQR_0, and outputs the frequency response FQR_0 to the collection means 41.

[0057] The estimation means 43 also reads out the time series data TSD_1(t) from the storage means 42, inputs the read out time series data TSD_1(t) to the learning device 44, and outputs N usability prediction results U1(t) to U N (t) U1(t)~U N Each of (t) is the result of predicting whether one reflective element 21 is usable or not.

[0058] Furthermore, the estimation means 43 reads out the time series data TSD_2(t) from the storage means 42, inputs the read out time series data TSD_2(t) to the learning device 45, and outputs N achievable rates AR1 to AR N receive.

[0059] The learning device 44 receives the time series data TSD_1(t) from the estimation means 43, and based on the received time series data TSD_1(t), calculates N usability prediction results U1(t) to U N (t) and the predicted N usability prediction results U1(t)~U N (t) is output to the estimation means 43.

[0060] The learning device 45 receives the time series data TSD_2(t) from the estimation means 43, and calculates N achievable rates AR1 to AR2 based on the received time series data TSD_2(t) by a method to be described later. N and the predicted N achievable rates AR1 to AR N is output to the estimation means 43.

[0061] Frequency response estimation Assuming that there is no fluctuation in the propagation path while observing the received signal y(t) (assuming that the frequency response H does not have (t)), and since beamforming is not performed in transmitter 1, the transmission weight W in transmitter 1 is omitted, the received signal y(t) can be expressed by the following equation.

[0062]

number

[0063] In equation (1), Θ is a matrix indicating N reflection coefficients of N reflecting elements 21, and n(t) represents white noise.

[0064] The received signal y(t) is T The observed M T By arranging the received signals y(t), the following equation is obtained:

[0065]

number

[0066] Then, in equation (2), multiplying Y(t) by the inverse matrix of X(t) from the right gives the following equation:

[0067]

number

[0068] In the last line of equation (3), H is the frequency response of the direct link H D and the frequency response H in the indirect link C and [H D H C ], and Ω is made up of a matrix in which an identity matrix I and Θ (the reflection coefficient matrix of N reflecting elements 21) are arranged vertically.

[0069] Then, when the modes of the N reflecting elements 21 are switched (that is, when the patterns of the N reflection coefficients of the N reflecting elements 21 are switched), the following equation is obtained.

[0070]

number

[0071] However, in equation (4), t = iM T As a result, t satisfies the relationship between the periods PD_1 to PD_M shown in FIG. R In each of x(1)~x(M T ) represents the timing of transmission, and t=iM T "i" in represents Ω(1) to Ω(N) shown in FIG.

[0072] When N G(i) values ​​in equation (4) are observed and the N G(i) values ​​are arranged, the following equation is obtained.

[0073]

number

[0074] Then, multiplying g in equation (5) by the inverse matrix of R from the right side gives the following equation:

[0075]

number

[0076] According to equation (5), R is made up of [Ω(1) Ω(N)], which is known because it is made up of N modes of reflection coefficients (i.e., patterns of N reflection coefficients) of N reflecting elements 21. Therefore, the inverse matrix of R can be calculated.

[0077] Also, G(t) = Y(t)X in equation (3) -1g in equation (6) can be calculated using (t), Y(t) is a signal received by the receiver 3 when the pilot signal is received, so it is known, and X(t) is a signal transmitted by the transmitter 1, so it is known. As a result, "gR" in equation (6) can be calculated using -1 " can be calculated.

[0078] Furthermore, "N" in equation (6) can be ignored when the amount of variation in the propagation path is sufficiently larger than the noise power (when the signal-to-noise ratio is high). When the amount of variation in the propagation path is not sufficiently larger than the noise power, the noise level in propagation path estimation can be suppressed to a predetermined value or less by transmitting the same pilot signal multiple times in succession and averaging the received signals.

[0079] Therefore, gR -1 and N.R. -1 By calculating the frequency response H D ,H C can be calculated independently. Frequency response H D is the frequency response of the direct link between transmitter 1 and receiver 3, and the frequency response H C is the frequency response of the indirect link between transmitter 1 and receiver 3.

[0080] As a result, the frequency response [H D H C ] can be estimated independently.

[0081] When the estimation means 43 of the estimation device 4 receives the received signal y(t), it calculates the frequency response [H D H C ](=FQR_0) is estimated.

[0082] [Collection of time-series data] Fig. 3 is a conceptual diagram of the time series data TSD_1(t), in which "I" represents an idle state and "B" represents a busy state.

[0083] Referring to FIG. 3, each of the idle state I and the busy state B is collected in units of slots SLT.

[0084] The collecting means 41 collects time series data TSD_1(t) for the following (A) or (B). (A) Collection of time series data TSD_1(t) when one or more of the N reflecting elements 21 are simultaneously turned OFF (B) N reflecting elements 21 are classified into clusters of a certain number (for example, 9), and the ON / OFF state of the reflecting elements is controlled in cluster units to collect time series data TSD_1(t). In (B), the collection means 41 selects, for all clusters, the reflective element located in the center of the arrangement positions of a certain number (two or more) of reflective elements classified into one cluster as a representative reflective element, and collects time series data indicating the ON / OFF state of the selected representative reflective element to collect time series data TSD_1(t).

[0085] Regarding the selection of the representative reflective element, the representative reflective element may be selected by a method other than the method of selecting the reflective element located at the center of the arrangement positions of a certain number of reflective elements classified into one cluster as the representative reflective element. For example, any reflective element from a certain number of reflective elements classified into one cluster may be selected as the representative reflective element. Any method for selecting a representative reflective element may be used.

[0086] In (A), when collecting time series data TSD_1(t), the collection means 41 receives from the control means 223 of the IRS controller 22 a period ON_period / period OFF_period_n(t) which is a time series arrangement of a period OFF_period in which one or more of the N reflecting elements 21 are turned OFF and a period ON_period in which one or more of the N reflecting elements 21 other than the one or more reflecting elements 21 are turned ON.

[0087] Then, the collection means 41 associates "I" with slot SLT_1 based on the period OFF_period, and associates "I" with slot SLT_2 based on the period ON_period, and thereafter similarly generates the time series data TSD_1(t) shown in Figure 3.

[0088] In this case, the collection means 41 detects that the reflecting elements that are turned OFF (one or more reflecting elements 21) are in an idle state I in slot SLT_1, and detects that the reflecting elements that are turned ON (reflecting elements other than one or more reflecting elements 21) are in a busy state B in slot SLT_2, and similarly detects that the reflecting elements that are turned OFF (one or more reflecting elements 21) are in an idle state I in slots SLT_3, SLT_5, SLT_6, and SLT_9, and detects that the reflecting elements that are turned ON (reflecting elements other than one or more reflecting elements 21) are in a busy state B in slots SLT_4, SLT_7, and SLT_8, thereby generating time series data TSD_1(t).

[0089] As a result, the time series of the idle state I and the busy state B can be detected for each of the N reflecting elements 21, and the time series data TSD_1(t) is made up of N B / I1(t) to B / I2(t) that indicate the time series of the idle state I and the busy state B for each of the N reflecting elements 21. N It consists of (t).

[0090] Then, the collecting means 41 stores the generated time series data TSD_1(t) in the storage means .

[0091] Furthermore, when collecting time series data TSD_1(t) in (B), the collection means 41 detects that in slot SLT_1, a cluster including a representative reflecting element that is turned OFF is in an idle state I, and in slot SLT_2, a cluster including a representative reflecting element that is turned ON is in a busy state B. Similarly, in slots SLT_3, SLT_5, SLT_6, and SLT_9, the collection means 41 detects that in slots SLT_3, SLT_5, SLT_6, and SLT_9, a cluster including a representative reflecting element that is turned ON is in an idle state I, and in slots SLT_4, SLT_7, and SLT_8, the collection means 41 detects that in slots SLT_4, SLT_7, and SLT_8, a cluster including a representative reflecting element that is turned ON is in a busy state B, thereby generating time series data TSD_1(t).

[0092] Then, the collecting means 41 stores the generated time series data TSD_1(t) in the storage means .

[0093] As long as a signal is being transmitted from the transmitter 1 to the receiver 3 via the IRS 2, the collection means 41 receives the period ON_period / period OFF_period_n(t) from the control means 223 of the IRS controller 22, and therefore continues to associate "I" or "B" with each slot SLT using the method described above.

[0094] As a result, the time series data TSD_1(t) represents the history of the idle state I and the busy state B (idle state I / busy state B).

[0095] 4 is a conceptual diagram of the time series data TSD_2(t). The collecting means 41 receives the frequency response matrix H(t) from the estimating means 43 and calculates the achievable rate R by the following method.

[0096] The N reflection coefficients θ and transmission weights W of the N reflecting elements are optimized by the following equations:

[0097]

number

[0098] In equation (7), γ mR is M R is the signal to interference and noise ratio (SINR) of the antenna, which is a function of the reflection coefficient θ and the transmit weight W. Also, ω mR is a parameter that indicates the priority of each receiving antenna. In particular, if no priority is given to the antennas, ω1 = ω2 = = ω MR =1 / M R This can be considered.

[0099] Equation (7) represents the determination of the reflection coefficient θ and the transmission weight W when the right-hand side is maximized.

[0100] And M R M antennas R The transmission rate is expressed as log2(1+γ mR ) is calculated as

[0101] The transmitter 1 multiplies (pre-codes) the transmission signal of each antenna by a coefficient matrix based on the information of the channel matrix (=frequency response matrix H(t)) acquired from the estimation device 4, so that the channels of the antenna pair become independent.

[0102] Therefore, the transmitter 1 determines the coefficient matrix by the following equation.

[0103]

number

[0104] In equation (8), matrix T is a transmission weight matrix, which corresponds to the above-mentioned transmission weight W. Furthermore, matrix A is a K×K diagonal matrix, where K is the number of streams, and p1, . . . , p K are the transmission powers allocated to streams 1 to K, respectively.

[0105] When the transmitted signal s is multiplied by the matrices T and A, the received signal y is expressed by the following equation.

[0106]

number

[0107] T to the received signal y H H H Multiplying by gives the following equation:

[0108]

number

[0109] ∧A in the last line of equation (10) is a diagonal matrix. Each diagonal element of ∧A (M R The result of dividing the noise power (or estimated noise power) of each antenna is γ1, γ2, . . . , γ MR This becomes:

[0110] Therefore, the collecting means 41 collects each diagonal element of ∧A (M R The result of dividing the noise power (or estimated noise power) of each antenna is γ mR Substituting log2(1+γ mR ) is calculated as γ1,γ2,...,γ MR Execute all of M R M antennas R Calculate the transmission rate.

[0111] And the collecting means 41 is M R The sum of the transmission rates is the achievable rate R.

[0112] Multiplying (precoding) the transmission signal of each antenna by the coefficient matrix G shown in equation (8) makes the propagation paths of the antenna pair independent, so M R Calculating the transmission rates of M R The propagation paths to the antennas are divided into M independent R Transform it into M propagation paths, andR M propagation paths R This corresponds to calculating the transmission rate.

[0113] The collecting means 41 collects the time series data TSD_2(t) shown in FIG. 4 by plotting the achievable rate R against the time t at which the achievable rate R is calculated.

[0114] The time series data TSD_2(t) is the time series data TSD_2(t) T Since this data is data collected of the achievable rate R when one or more reflecting elements 21 that are turned on when transmitting a signal using N antennas (when N reflecting elements 21 are clustered, the on / off patterns of all reflecting elements 21 in the cluster are assumed to be the same), the collection means 41 generates time series data TSD_2_n(t) for all patterns of one or more reflecting elements 21 that are turned on. As a result, the time series data TSD_2(t) is made up of E pieces of time series data TSD_2_1(t) to TSD_2_E(t), where E is the number of patterns of one or more reflecting elements 21 that are turned on.

[0115] E is the number of reflection patterns of one or more reflective elements 21 when one or more reflective elements 21 of the N reflective elements 21 are used (ON), and is an integer equal to or greater than 2. When reflective element 21_1 of the N reflective elements 21 is used, the reflection pattern of the one or more reflective elements 21 is, for example, [reflective element 21_1=ON, reflective element 21_2: OFF, reflective element 21_3: OFF, . . . , reflective element 21_N: OFF], and when reflective elements 21_1 and 21_2 of the N reflective elements 21 are used, the reflection pattern is [reflective element 21_1=ON, reflective element 21_2: ON, reflective element 21_3: OFF, . . . , reflective element 21_N: OFF]. The same applies to the patterns of one or more reflective elements 21 when one or more reflective elements 21 other than reflective elements 21_1 and 21_2 of the N reflective elements 21 are used simultaneously.

[0116] As long as a signal is being transmitted from the transmitter 1 to the receiver 3 via the IRS 2, the collection means 41 continues to calculate the achievable rate R using the method described above and plot the calculated achievable rate R against the calculation time t of the achievable rate R.

[0117] As a result, the time series data TSD_2(t) represents the history of the achievable rate R.

[0118] [Prediction of usability results] Fig. 5 is a schematic diagram of a machine learning device that constitutes the learning device 44 shown in Fig. 2. Referring to Fig. 5, the learning device 44 includes N probabilistic neural networks PNN. SLT _1~PNN SLT It consists of _N.

[0119] As described above, the time series data TSD_1(t) is composed of N B / I1(t) to B / I N (t), the learner 44 is composed of N probabilistic neural networks PNN SLT _1~PNN SLT It consists of _N.

[0120] N probabilistic neural networks PNN SLT _1~PNN SLT _N are the time series data TSD_1_1(t)~TSD_1_N(t) (=B / I1(t)~B / I N (t)).

[0121] Here, each of the time series data TSD_1_1(t) to TSD_1_N(t) is made up of data in which the period OFF_period of the idle state I and the period ON_period of the busy state B for one of the N reflecting elements 21 are arranged in time series.

[0122] And N probabilistic neural networks (PNN) SLT _1~PNN SLT _N are the time series data TSD_1_1(t)~TSD_1_N(t) (=B / I1(t)~B / IN Based on the N reflecting elements 21, it is predicted whether or not the reflecting element 21 can be used in the upcoming slot SLT, and the predicted results are used / unused prediction results U1(t) to U N (t) is output to the estimation means 43. In this case, the usability prediction results U1(t) to U N Each of (t) consists of "0" which indicates that one reflective element 21 is in the idle state I, or "1" which indicates that one reflective element 21 is not in the idle state I.

[0123] Figure 6 shows the probabilistic neural network (PNN) shown in Figure 5. SLT 6 is a schematic diagram of a probabilistic neural network PNN. SLT _1 comprises an input layer, a hidden layer, a summation layer, and a decision layer.

[0124] The input layer consists of Q neurons (Q is an integer greater than or equal to 2). The hidden layer consists of M groups (M is an integer greater than or equal to 2), each of which consists of Q neurons. The sum layer consists of M neurons.

[0125] Each of the Q neurons in the input layer is connected to each of the Q×M neurons in the hidden layer.

[0126] In the hidden layer, each group of Q neurons x 1,1 ~x 1,Q is connected to one neuron in the sum layer.

[0127] The M neurons in the summation layer are connected to one neuron in the decision layer.

[0128] The input layer normalizes the input data and outputs the normalized data to the hidden layer. In this case, each of the Q neurons in the input layer outputs normalized data to each of the Q×M neurons in the hidden layer.

[0129] The hidden layer calculates the Euclidean distance between the time series data TSD_1_COL(t) collected by the collection means 41 and the time series data TSD_1_SAV(t) (time series data TSD_1(t) collected in the learning stage) stored in the hidden layer. Then, the hidden layer outputs the calculated Euclidean distance to the summation layer. In this case, in the hidden layer, Q neurons x 1,1 ~x 1,Q outputs the Euclidean distance to one neuron in the sum layer that is associated with the class 1 group. Similarly, Q neurons x 1,1 ~x 1,Q outputs the Euclidean distance to one neuron in the summation layer that corresponds to the class M group.

[0130] The summation layer sums the contributions of each class c to produce a probability given by:

[0131]

number

[0132] In equation (11), x is the input vector, and x c,j is the jth training vector, and N R is the total number of training vectors, and σ c is the standard deviation of class c. Also, f c is that the collected vector x is a predictor matrix P SLT r Class c of the data (time series data TSD_1(t) in the learning stage) stored in (the prediction matrix stored in the summation layer) 1 learn represents the probability of matching.

[0133] The summation layer then outputs the probability generated by equation (11) to the decision layer. In this case, each of the M neurons in the summation layer outputs a probability generated by equation (11) to one neuron in the decision layer.

[0134] The decision layer finds the maximum probability among the M probabilities received from the M neurons in the summation layer, and determines the class c when the maximum probability is obtained. COL_1 Select and output Class C COL_1 is a class classified according to the number of idle states I in the time series data TSD_1(t), and includes the number of idle states I in the time series data TSD_1(t). COL_1 constitute the "first class."

[0135] Prediction matrix P SLT r The data stored in (time series data TSD_1(t) in the learning stage) is the time series data TSD_1_learn(t) acquired in the learning stage of the probabilistic neural network PNN_1.

[0136] In the learning stage, the control means 223 of the IRS controller 22 instructs the transmitter 1 to transmit a training signal to the IRS 2 periodically (for example, once every 1 to 10 minutes).

[0137] The transmitter 1 transmits a training signal to the IRS 2 in response to an instruction from the control means 223. In this case, the transmitter 1 transmits the training signal to the IRS 2 in response to an instruction from the control means 223. T Use antennas to transmit training signals to IRS2.

[0138] In the IRS2, the control means 223 controls the reflection beamforming vectors φ of the N reflecting elements 21 so as to turn on one or more of the N reflecting elements 21 that are in the OFF state, and the N reflecting elements 21 control the reflection angle and phase of the radio wave according to the control from the control means 223 to reflect the training signal in the direction of the receiver 3. In this case, the control means 223 holds a codebook, and uses the codebook to control the reflection patterns of the N reflecting elements 21 into V (V is an integer equal to or greater than 2) reflection patterns by V reflection beamforming vectors φ1 to φ V and the selected V reflected beamforming vectors φ1 to φ V One reflected beamforming vector φn (n is any one of 1 to V) to switch the reflected beam forming vectors of N reflecting elements 21 to V reflected beam forming vectors φ1 to φ V Execute all of the above.

[0139] Then, V reflected beamforming vectors φ1 to φ V Each of the V reflected beam forming vectors φ1 to φ2 is a reflected beam forming vector in which one or more of the N reflecting elements 21 are simultaneously turned on. V to one reflected beamforming vector φ n The selecting is performed according to a predetermined order.

[0140] Therefore, the control means 223 selects V reflected beamforming vectors φ1 to φ from the codebook. V Select .

[0141] The control means 223 of the IRS2 then calculates V reflected beamforming vectors φ1 to φ V is used to obtain data D_ON / OFF_t consisting of a time series of ON / OFF for the N reflecting elements 21 when the reflection angle and phase of the radio waves by the N reflecting elements 21 are switched, and based on the obtained data D_ON / OFF_t, time series data D_I / B_1(t) to D_I / B_N(t) consisting of idle state I / busy state B for the N reflecting elements 21 is obtained.

[0142] Then, the control means 223 of the IRS 2 outputs the acquired time series data D_I / B_1(t) to D_I / B_N(t) to the estimation means 43 of the estimation device 4 as time series data TSD_1_learn(t).

[0143] The estimation means 43 of the estimation device 4 receives the time series data TSD_1_learn(t) (= time series data D_I / B_1(t) to D_I / B_N(t)) from the control means 223, detects usable reflecting elements among the N reflecting elements 21 based on the received time series data TSD_1_learn(t) (= time series data D_I / B_1(t) to D_I / B_N(t)), and determines the number of usable reflecting elements according to the detected number of usable reflecting elements. R (n R is an integer greater than or equal to 2) of classes c 1 learn_1 ~c 1 learn_nR Classify into n R Class C 1 learn_1 ~c 1 learn_nR respectively as PNN SLT Prediction matrix P in _1 SLT r n R Store in n rows. R Class C 1 learn_1 ~c 1 learn_nR Each of these constitutes a "first learning stage class."

[0144] During the learning phase, the above operations are performed periodically, so that the PNN SLT Prediction matrix P in _1 SLT r n R n rows stored R Class C 1 learn_1 ~c 1 learn_nR will be updated from time to time.

[0145] FIG. 7 shows the prediction matrix P SLT r 7 is a conceptual diagram showing the time series data TSD_1(t). Referring to Fig. 7, the time series data TSD_1(t) includes "■" and "□". "■" represents a busy state B, and "□" represents an idle state I. Each of the "■" and "□" is placed in one slot SLT.

[0146] Prediction matrix P SLT r is n R row n C The sliding window SDW1 consists of a matrix of columns. C .ts(s).

[0147] The sliding wind vector s is the vector of the sliding wind SDW1. n is the prediction matrix P SLT r The sliding wind SDW1 forms one row for wind span n C Moving to the right at .ts(s) (see the arrow in Figure 7), one new sample consisting of idle state I / busy state B (one sample of the time series data TSD_1(t)) is added to the prediction matrix P SLT r The sliding window SDW1 is moved to the right, and the sliding window vector s n Using the time series data TSD_1(t), one sample of the time series data TSD_1(t) is estimated using the prediction matrix P SLT r and drop it into one row of the time series data TSD_1(t) R The samples are then converted into a prediction matrix P SLT r n R Drop it onto this line.

[0148] n R Each row T of the ×1 target matrix contains a positive integer indicating an achievable usable class c of reflective elements (1≦T≦K), where K is the total number of achievable usable classes c of reflective elements and is an integer equal to or greater than 1.

[0149] Prediction matrix P SLT r is the sample dropped in one row using the sliding window SDW1 (sliding window vector s n ~s n-nC+1Class c classified according to idle state I of samples (samples dropped by COL_1 Class c classified according to the idle state I of the time series data (time series data TSD_1_learn(t) obtained in the learning stage) stored in one row 1 learn The probability of matching n R Execute for rows.

[0150] And the target matrix is ​​the prediction matrix P SLT r n R n received from rows R The positive integer contained in the row T that received the maximum probability among the probabilities (a positive integer indicating an idle state I or a busy state B) is output.

[0151] In this way, probabilistic neural networks (PNNs) SLT _1 is the prediction matrix P for one reflecting element. SLT r is used to predict the idle state I or busy state B that is closest to the pattern history (time series data TSD_1(t)).

[0152] As a result, the probabilistic neural network (PNN) shown in Figure 5 was obtained. SLT _1 is a function that when one time series data TSD_1(t) is input, the input time series data TSD_1(t) is converted into a prediction matrix P SLT r Class c of the data (time series data TSD_1_learn(t)) stored in (the prediction matrix held in the summation layer) learn The class c when the maximum probability of matching is obtained max (That is, the availability prediction result U1(t) consisting of the idle state I or the busy state B) is output.

[0153] The probabilistic neural network (PNN) shown in Figure 5 SLT _2~PNN SLT Each of the N is a probabilistic neural network PNN shown in FIG. SLTIt has the same structure as _1. Therefore, it is a probabilistic neural network (PNN). SLT _2~PNN SLT _N are the time series data B / I2(t)~B / I N When (t) is input, the usability prediction results U2(t)~U for N-1 reflecting elements are N Output (t).

[0154] Therefore, it is possible to predict which of the N reflecting elements 21 can be used.

[0155] As a result, a communication schedule can be created using the predicted available reflecting elements.

[0156] FIG. 8 shows the prediction matrix P SLT r 8 is a schematic diagram showing a specific example of the above. In Fig. 8, "0" represents the idle state I, and "1" represents not being in the idle state I.

[0157] Referring to FIG. 8, the sliding window vector s shown in FIG. n Using the time series data TSD_1(t), one sample of the time series data TSD_1(t) is estimated using the prediction matrix P SLT r Drop it into one row of the time series data TSD_1(t) and add 100 (=n R ) samples into the prediction matrix P SLT r 100 (=n R ) rows.

[0158] This allows us to calculate the time t PDC The past idle / busy pattern I / B_PTN is predicted by the prediction matrix P SLT r 100 (=n R ) rows.

[0159] The time t at which the availability of one reflective element 21 is predicted PDC In this case, the availability of one reflecting element 21 is predicted.

[0160] As a result, "0" is stored in the "prediction result of usability" for sequences 1 to 3, and "1" is stored in the "prediction result of usability" for sequences 1 to 3.

[0161] For sequences 1 to 3, the predicted availability is "0", so at time t PDC In the slot SLT after , one reflective element is expected to be idle.

[0162] On the other hand, for sequence 100, the predicted result of availability is "1", so at time t PDC In the slot SLT after , one reflective element is predicted not to be idle.

[0163] Since each of the prediction results of usability in sequences 1 to 100 is expressed as a probability, the prediction result of usability with the highest probability is ultimately selected from the 100 prediction results of usability in sequences 1 to 100.

[0164] Therefore, when the predicted result of the finally selected usability is the idle state (0), one reflective element is predicted to be usable, and when the predicted result of the finally selected usability is not the idle state (0), one reflective element is predicted to be unusable.

[0165] Therefore, N probabilistic neural networks (PNN) shown in Figure 5 SLT _1~PNN SLT _N are the reflection elements 21 at time t PDC The estimation means 43 then predicts whether the packet can be used in the slot SLT that follows the packet.

[0166] [Achievable Rate AR Prediction] Fig. 9 is a schematic diagram of a machine learning device constituting the learning device 45 shown in Fig. 2. Referring to Fig. 9, the learning device 45 includes E probabilistic neural networks PNNs. AR _1~PNN AR It consists of _E.

[0167] E probabilistic neural networks PNN AR _1~PNN AR _E are the time series data TSD_2_1(AR1(t)_ COL )~TSD_2_E(AR E (t)_ COL ) from the estimation means 43.

[0168] And E probabilistic neural networks (PNNs) AR _1~PNN AR _E are the time series data TSD_2_1(AR1(t)_ COL )~TSD_2_E(AR E (t)_ COL ), the achievable rates AR1(t) to AR E (t) and estimate the estimated achievable rate AR1(t)~AR E (t) is output to the estimation means 43.

[0169] E probabilistic neural networks PNN AR _1~PNN AR Each of the _E is a probabilistic neural network PNN shown in Figure 6. SLT It has the same structure as _1.

[0170] E probabilistic neural networks PNN AR _1~PNN AR In each of the E models, the input layer normalizes the input data and outputs the normalized data to the hidden layer. In this case, each of the Q neurons in the input layer outputs normalized data to each of the Q×M neurons in the hidden layer.

[0171] The hidden layer calculates the Euclidean distance between the collected time series data TSD_2_COL(t) and the time series data TSD_2_SAV(t) stored in the hidden layer. The hidden layer then outputs the calculated Euclidean distance to the summation layer. In this case, in the hidden layer, Q neurons x of the class 1 group are 1,1 ~x 1,Q outputs the Euclidean distance to one neuron in the sum layer that is associated with the class 1 group. 1,1 ~x 1,Q outputs the Euclidean distance to one neuron in the summation layer that corresponds to the class M group.

[0172] The summation layer sums the contributions of each class c to generate the probability expressed by equation (11). When a probabilistic neural network (PNN) is used to predict the achievable rate (AR), f in equation (11) c is that the collected vector x is a predictor matrix P AR r Class c of the data (time series data TSD_2(t)) stored in (the prediction matrix held in the summation layer) 2 learn represents the probability of matching.

[0173] The summation layer then outputs the probability generated by equation (11) to the decision layer. In this case, each of the M neurons in the summation layer outputs a probability generated by equation (11) to one neuron in the decision layer.

[0174] The decision layer finds the maximum probability among the M probabilities received from the M neurons in the summation layer, and determines the class c when the maximum probability is obtained. 2 learn Select and output.

[0175] Prediction matrix P AR r The data stored in (time series data TSD_2(t)) is processed by E probabilistic neural networks PNN. AR _1~PNN AR_E is the time series data TSD_2_learn(t) obtained at each learning stage.

[0176] In the learning stage, the control means 223 of the IRS controller 22 instructs the transmitter 1 to transmit a training signal to the IRS 2 periodically (for example, once every 1 to 10 minutes).

[0177] The transmitter 1 transmits a training signal to the IRS 2 in response to an instruction from the control means 223. In this case, the transmitter 1 transmits the training signal to the IRS 2 in response to an instruction from the control means 223. T Use antennas to transmit training signals to IRS2.

[0178] In the IRS2, the control means 223 uses the reflection beamforming vector φ to control the reflection angle and phase of the radio waves by the N reflecting elements 21, and the N reflecting elements 21 change the reflection angle and phase of the radio waves according to the control from the control means 223 to reflect the training signal in the direction of the receiver 3. In this case, the control means 223 selects V reflection beamforming vectors φ1 to φ from the codebook. V and the selected V reflected beamforming vectors φ1 to φ V One reflected beamforming vector φ m (m is any one of 1 to V) to switch the reflected beam forming vectors of the N reflecting elements 21 to V reflected beam forming vectors φ1 to φ V Then, V reflected beamforming vectors φ1 to φ V are reflected beam forming vectors for controlling the N reflecting elements 21 so that one or more of the N reflecting elements 21 are turned ON and the other reflecting elements 21 are turned OFF. Also, V reflected beam forming vectors φ1 to φ V to one reflected beamforming vector φ m The selecting is performed according to a predetermined order.

[0179] Thereafter, the estimation means 43 calculates the reflected beamforming vectors of the N reflecting elements 21 into one reflected beamforming vector φ m When the switch is made to the diagonal elements of ∧A shown in equation (10) (M R The result of dividing the noise power (or estimated noise power) of each antenna is γ mR Substituting log2(1+γ mR ) is calculated as γ1,γ2,...,γ MR Execute all of M R M antennas R Calculate the transmission rate.

[0180] Then, the estimation means 43 calculates M R The sum of these transmission rates is calculated as the achievable rate R.

[0181] The estimation means 43 calculates the achievable rate R using V reflected beamforming vectors φ1 to φ V , and E achievable rates AR1 to AR E Calculate.

[0182] Then, the estimation means 43 estimates the achievable rates AR1 to AR E Depending on the value of AR, the achievable rate is n R Class C 2 learn_1 ~c 2 learn_nR Classify into n R Class C 2 learn_1 ~c 2 learn_nR respectively as PNN AR Prediction matrix P for m (m is 1 to E) AR r n R The PNN stores the data in rows. AR _1~PNN AR _Execute for all of E.

[0183] During the learning phase, the above operations are periodically performed, so that E probabilistic neural networks PNN are generated. AR _1~PNN AR Prediction matrix P for each of E AR r n R n rows stored R Class C 2 learn_1 ~c 2 learn_nR will be updated from time to time.

[0184] Figure 10 shows the prediction matrix P that predicts the achievable rate AR. AR r 10, the time series data TSD_2(t) is time series data collected by the collecting means 41.

[0185] Prediction matrix P AR r is n R row n C The sliding window SDW2 consists of a matrix of columns. C .ts(s).

[0186] The sliding wind vector s is the vector of the sliding wind SDW2. n is the prediction matrix P AR r The sliding wind SDW2 forms one row for wind span n C Moving to the right at .ts(s) (see the arrow in Figure 10), one sample of one new achievable rate AR (one sample of the time series data TSD_2(t)) is added to the prediction matrix P AR r The sliding window vector s is dropped into one row of the sliding window SDW2. n Using the time series data TSD_2(t), one sample of the time series data TSD_2(t) is predicted using the prediction matrix P AR r and drop it into one row of the time series data TSD_2(t) R The samples are then converted into a prediction matrix PAR r n R Drop it onto this line.

[0187] n R Each row of the × 1 target matrix T contains a class c of achievable rates AR. 2 learn contains a positive integer indicating (1≦T≦K).

[0188] Prediction matrix P AR r is the sample dropped in one row using the sliding window SDW2 (sliding window vector s n ~s n-nC+1 Class c of samples dropped by COL_2 The class c of the time series data (time series data TSD_2_learn(t) obtained in the learning stage) is stored in one row. 2 learn The probability of matching n R Execute for each row. Note that class c COL_2 constitutes the "second class" and class c 2 learn constitute the "second learning stage class."

[0189] And the target matrix is ​​the prediction matrix P AR r n R n received from rows R The positive integer contained in the row T that received the maximum probability among the probabilities (the positive integer indicating the achievable rate AR) is output.

[0190] In this way, E probabilistic neural networks (PNN) AR _1~PNN AR Each of _E is a prediction matrix P AR r is used to find the maximum achievable rate AR that is closest to the pattern history (time series data TSD_2(t)).

[0191] As a result, E probabilistic neural networks (PNN) are generated.AR _1~PNN AR When one time series data TSD_2(t) is input, each of the time series data TSD_2(t) is classified into the class c COL_2 is the prediction matrix P AR r Class c of the data (time series data TSD_2_learn(t)) stored in (the prediction matrix held in the summation layer) 2 learn The class c when the maximum probability of matching is obtained 2 max (i.e., the achievable rate AR1(t) to AR E (t)).

[0192] Therefore, transmitter 1 is M T The achievable rate AR when transmitting a signal using this antenna can be predicted for one or more of the N reflecting elements 21 being utilized simultaneously.

[0193] The estimation means 43 includes E probabilistic neural networks PNN. AR _1~PNN AR E to E achievable rates AR1(t) PRD ~AR E (t)_ PRD The estimation means 43 then receives E achievable rates AR1(t)_ PRD ~AR E (t)_ PRD Based on this, the achievable rate AR1(t) of the reflective element 21 predicted to be usable in the future (one or more reflective elements 21 whose predicted usability is "0") is calculated. PRD ~AR e’ (t)_ PRD , and the detected achievable rate AR1(t)_ PRD ~AR e’ (t)_ PRD The maximum achievable rate AR of max and the maximum achievable rate AR max The reflected beamforming vector φ of the N reflecting elements 21 when maxHere, e' is in the range of 1≦e'≦E.

[0194] As a result, the maximum achievable rate AR max A communication schedule using the above can be created.

[0195] The maximum achievable rate (AR) max is one reflected beam forming vector φ in which one or more of the N reflecting elements 21 are simultaneously turned on. m V reflected beamforming vectors φ1 to φ V E achievable rates AR1 to AR2 estimated by switching over all of E Therefore, the maximum achievable rate among the E achievable rates AR1 to AR E Each of the rates is the achievable rate for the reflecting element being used among the N reflecting elements 21.

[0196] [Estimation of reflected beamforming vector] The codebook is designed by using the discrete Fourier transform of the reflection beamforming vector φ of the reflection element, and the codebook has a size of, for example, 32, 64, . . . , 1024. The size is a parameter.

[0197] The reflective elements other than the newly turned ON reflective element are set to "0".

[0198] A codebook designed using the Discrete Fourier Transform (DFT-M) H is the coherence of the azimuthal dimension. H The "M" in " represents the horizontal direction.

[0199] m H The th column element is expressed by the following formula:

[0200]

number

[0201] In equation (12), m H =1,2,...,M H where j represents the imaginary unit.

[0202] M H For the case where =4, the codebook DFT-M H is shown in Table 1.

[0203] [Table 1]

[0204] One or more of the N reflecting elements 21 that are utilized simultaneously may be used to calculate the maximum achievable rate R max is obtained, the estimation means 43 calculates the maximum achievable rate R max The reflected beamforming vector is estimated by selecting from the codebook a reflected beamforming vector (a reflected beamforming vector for one or more reflecting elements 21 used simultaneously) when

[0205] As described above, the estimation means 43 of the estimation device 4 receives the usability prediction results U1(t) to U2(t) for the N reflecting elements 21 from the learning device 44. N (t), and E achievable rates AR1(t) to AR E Receive (t).

[0206] Then, the estimation means 43 estimates the usability prediction results U1(t) to U N Based on (t), the estimation means 43 detects usable reflecting elements among the N reflecting elements 21. For example, the estimation means 43 detects two reflecting elements 21_1 and 21_2 as usable reflecting elements among the N reflecting elements 21.

[0207] Then, when the estimation means 43 detects the two reflecting elements 21_1 and 21_2, it obtains three combinations of "reflecting element 21_1 only," "reflecting element 21_2 only," and "reflecting element 21_1 and reflecting element 21_2" as combinations when using the two reflecting elements 21_1 and 21_2.

[0208] Then, the estimation means 43 estimates the achievable rate AR when using "only the reflecting element 21_1". 21_1 and the achievable rate AR when using only the reflective element 21_2. 21_2 and the achievable rate AR when using the "reflection element 21_1 and the reflection element 21_2". 21_1,21_2 and E achievable rates AR1(t)~AR E Detect from (t).

[0209] E achievable rates AR1(t)~AR E Since (t) is the achievable rate for all combinations when one or more of the N reflecting elements 21 are used, the estimation means 43 calculates the achievable rate AR 21_1 , achievable rate AR 21_2 and achievable rate AR 21_1,21_2 E achievable rates AR1(t)~AR E It can be detected from (t).

[0210] The estimation means 43 estimates the achievable rate AR 21_1 , achievable rate AR 21_2 and achievable rate AR 21_1,21_2 When detecting the achievable rate AR 21_1 , achievable rate AR 21_2 and achievable rate AR 21_1,21_2 The maximum achievable rate AR max Select the maximum achievable rate AR max The reflected beamforming vector φ when max is selected from the codebook.

[0211] Here, the estimation means 43 estimates the maximum achievable rate AR maxAs the achievable rate AR 21_1,21_2 shall be detected.

[0212] The estimation means 43 then estimates the reflecting elements 21_1 and 21_2 that are available among the N reflecting elements 21, the maximum achievable rate AR max (=AR 21_1,21_2 ), and the reflected beamforming vector φ max Create a communication schedule based on this.

[0213] Therefore, it is possible to create a schedule for using the reflecting elements by determining in advance whether there are enough reflecting elements available to actually ensure a sufficient transmission rate.

[0214] Fig. 11 is a flowchart for explaining the operation of the estimation device 4 shown in Fig. 2. Referring to Fig. 10, when the operation of the estimation device 4 is started, the estimation means 43 of the estimation device 4 calculates the frequency response matrix H C ,H D is estimated (step S1).

[0215] The collecting means 41 of the estimation device 4 receives the ON / OFF periods of the N reflecting elements detected by the control means 223 from the control means 223, and collects time series data TSD_1(t) indicating the time series of the ON / OFF periods of the N reflecting elements based on the received ON / OFF periods of the N reflecting elements (step S2). The collecting means 41 then stores the time series data TSD_1(t) in the storage means 42.

[0216] In this case, the control means 223 detects the period during which one reflecting element is ON and the period during which one reflecting element is OFF, and arranges the detected ON periods and OFF periods in time series to generate time series data for one reflecting element, performing this process for all N reflecting elements to generate time series data TSD_1(t). The control means 223 then outputs the time series data TSD_1(t) to the collection means 41, and the collection means 41 collects the time series data TSD_1(t) by receiving the time series data TSD_1(t) from the control means 223. The collection means 41 also stores the time series data TSD_1(t) in the storage means 42. Furthermore, the time series data TSD_1(t) consists of time series data D_I / B_1(t) to D_I / B_N(t) for the reflecting elements 21_1 to 21_N, respectively.

[0217] After step S2, the collecting means 41 calculates the achievable rate R using the frequency response matrix H(t) according to equations (7) to (10) and collects the time series data TSD_2(t) (step S3). Then, the collecting means 41 stores the time series data TSD_2(t) in the storage means 42.

[0218] Subsequently, the estimation means 43 reads the time series data TSD_1(t) from the storage means 42, and calculates the time series data TSD_1(t) based on the read time series data TSD_1(t) using a learning device 44 (PNN SLT _1~PNN SLT _N), the usable reflecting elements among the N reflecting elements 21 are estimated (step S4).

[0219] The estimation means 43 includes a learning device 45 (PNN AR _1~PNN AR _E), based on the time series data TSD_2(t), E achievable rates R1 to R2 when one or more of the N reflecting elements are used simultaneously are calculated. E is estimated (step S5).

[0220] Then, the estimation means 43 calculates the achievable rates R1 to R E The maximum achievable rate R maxThe reflected beamforming vector φ of N reflecting elements when max is acquired (step S6).

[0221] Then, the estimation means 43 estimates the number of usable reflecting elements among the N reflecting elements 21, the number of E achievable rates R1 to R2, and the number of usable reflecting elements among the N reflecting elements 21. E The maximum achievable rate R max and the reflected beamforming vector φ max Then, a communication schedule using IRS2 is created using the above (step S7). This completes the operation of the estimation device 4.

[0222] In the flowchart shown in FIG. 11, the maximum achievable rate R max The reflected beamforming vector φ of N reflecting elements when max (see step S6) This is to maximize the transmission rate in the downlink.

[0223] Therefore, the estimation means 43 estimates the reflected beamforming vector φ max A communication schedule is created using the above (see step S7).

[0224] FIG. 12 is a flowchart for explaining the detailed operation of step S4 shown in FIG.

[0225] Referring to FIG. 12, after step S3 of FIG. 11, the probabilistic neural network PNN of the learning device 44 SLT _e (e is any of 1 to E) receives the time series data D_I / B_e(t) from the estimation means 43 (step S41).

[0226] and probabilistic neural networks (PNNs). SLT _e is the sliding wind vector s n n extracted from the time series data D_I / B_e(t) using C A sample consisting of data is expressed as a prediction matrix P SLT rBy repeatedly dropping the data into one row of the predicted matrix P SLT r n R n in each of the rows C samples are dropped (step S42).

[0227] Then, the probabilistic neural network (PNN) SLT Prediction matrix P of _e SLT r is the prediction matrix P SLT r A sample class of one row of COL_1 is the prediction matrix P SLT r The class c of the training sample stored in one row of 1 learn The probability of matching n R This is executed for all of the rows (step S43).

[0228] Then, the probabilistic neural network (PNN) SLT The target matrix T of _n is the prediction matrix P SLT r from n R The positive integer included in the row of the target matrix T that has received the maximum probability out of the probabilities (the positive integer indicating the prediction result of usability) is output to the estimation means 43 (step S44). After step S44, the operation of the estimation device 4 proceeds to step S5 in FIG.

[0229] In step S44 of FIG. R If there are multiple maximum probabilities among the probabilities, the target matrix T outputs to the estimation means 43 a positive integer (a positive integer indicating the prediction result of usability) contained in the row of the target matrix T that has received any one of the multiple maximum probabilities.

[0230] Also, probabilistic neural networks (PNNs) SLT _1~PNN SLT_N executes the flowchart shown in FIG. 12 in parallel, and outputs U1(t) to U N (t) is output to the estimation means 43. As a result, usable reflecting elements among the N reflecting elements 21 are calculated by the probabilistic neural network PNN SLT _1~PNN SLT It is estimated by _N.

[0231] FIG. 13 is a flowchart for explaining the detailed operation of step S5 shown in FIG.

[0232] Referring to FIG. 13, after step S4 of FIG. 11, the probabilistic neural network PNN of the learning device 45 AR _e (e is any of 1 to E) receives the time series data TSD_2_e(t) from the estimation means 43 (step S51).

[0233] and probabilistic neural networks (PNNs). AR _e is the sliding wind vector s n n extracted from the time series data TSD_2_e(t) using C A sample consisting of data is expressed as a prediction matrix P AR r By repeatedly dropping the data into one row of the predicted matrix P AR r n R n in each of the rows C samples are dropped (step S52).

[0234] Then, the probabilistic neural network (PNN) AR Prediction matrix P of _e AR r is the prediction matrix P AR r A sample class of one row of COL_2 is the prediction matrix P AR r The class c of the training sample stored in one row of 2 learn The probability of matching nR This is executed for all of the rows (step S53).

[0235] Then, the probabilistic neural network (PNN) AR The target matrix T of _e is the prediction matrix P AR r from n R The positive integer (positive integer indicating the achievable rate AR) included in the row of the target matrix T that has received the maximum probability out of the probabilities is output to the estimation means 43 (step S54). After step S54, the operation of the estimation device 4 proceeds to step S6 in FIG.

[0236] In step S54 of FIG. R If there are multiple maximum probabilities among the probabilities, the target matrix T outputs to the estimation means 43 a positive integer (a positive integer indicating the achievable rate AR) contained in the row of the target matrix T that has received any one of the multiple maximum probabilities.

[0237] Also, probabilistic neural networks (PNNs) AR _1~PNN AR _E executes the flowchart shown in FIG. 13 in parallel to obtain the achievable rates AR1(t) to AR E (t) of the N reflecting elements 21. E (t) is a probabilistic neural network (PNN) AR _1~PNN AR It is estimated by _E.

[0238] According to the flowchart shown in FIG. 11 (including the flowcharts shown in FIGS. 12 and 13), the estimation means 43 estimates the number of available reflecting elements among the N reflecting elements 21 and the number of E achievable rates AR1 to AR2. E and E achievable rates AR1 to AR E The maximum achievable rate AR max The reflected beamforming vector φ of the N reflecting elements 21 when max Get.

[0239] The estimation means 43 then estimates the number of usable reflecting elements among the N reflecting elements 21, the number of E achievable rates AR1 to AR2, and the number of usable reflecting elements 21. E The maximum achievable rate AR max , and the reflected beamforming vectors φ of the N reflecting elements 21 max Use this to create a communication schedule using IRS2.

[0240] Then, the estimation means 43 estimates the number of usable reflecting elements among the N reflecting elements 21, the number of E achievable rates R1 to R2, and the number of usable reflecting elements among the N reflecting elements 21. E The maximum achievable rate AR max , the reflected beamforming vector φ of the N reflecting elements 21 max And output the communication schedule to the control means 223.

[0241] The control means 223 determines the number of available reflecting elements among the N reflecting elements 21, the number of E achievable rates AR1 to AR2, and E The maximum achievable rate AR max , the reflected beamforming vector φ of the N reflecting elements 21 max and the communication schedule from the estimation means 43, the available reflecting elements among the N reflecting elements 21, the E achievable rates AR1 to AR2, E The maximum achievable rate AR max and a communication schedule to the transmitter 1, and the reflected beamforming vectors φ of the N reflecting elements 21. max Hold.

[0242] The transmitter 1 selects the usable reflecting elements among the N reflecting elements 21, and the E achievable rates AR1 to AR2. E The maximum achievable rate AR max and receives a communication schedule from the control means 223 of the IRS2.

[0243] The transmitter 1 then selects N usable reflecting elements out of the N reflecting elements 21 and E achievable rates AR1 to AR2. EThe maximum achievable rate AR max , and the resources to be used for future communication (the available reflective elements among the N reflective elements 21 and the maximum achievable rate AR max ) is present or not.

[0244] When the transmitter 1 determines that there are resources available for future transmission, it transmits a communication notification to the control means 223 of the IRS 2 and starts communication.

[0245] After receiving the communication notification from the transmitter 1, the control means 223 of the IRS 2 controls the reflected beamforming vector φ max Based on the reflected beamforming vector φ max The pattern of the N reflecting elements 21 is controlled to match the reflection pattern when the reflecting element 21 is used.

[0246] Therefore, the usable reflecting elements among the N reflecting elements 21 and the E achievable rates AR1 to AR2 are E The maximum achievable rate AR max By estimating the N reflecting elements 21, the available reflecting elements and the maximum achievable rate AR max Therefore, it is possible to create a schedule for using the reflecting elements by determining in advance whether enough reflecting elements are available to ensure a sufficient transmission rate.

[0247] Furthermore, the estimation means 43 calculates the time series data B / I1(t) to B / I N (t) are respectively the probabilistic neural network PNN SLT _1~PNN SLT Input to _N and generate a probabilistic neural network (PNN) SLT _1~PNN SLT _N execute the flowchart shown in FIG. 12 in parallel, and obtain the usability prediction results U1(t) to U N Outputting (t) to the estimation means 43 means that the estimation means 43 outputs the first idle state / busy state data to one first machine learning machine (PNN SLT_1~PNN SLT _N), and executes a first process of receiving from one first machine learning device a usability prediction result corresponding to the maximum probability among the probabilities that a first class classified according to the idle state of the first idle state / busy state data matches a first learning stage class, which is a class classified according to the idle state of the first idle state / busy state data of the learning stage acquired in the learning stage, and estimating usability prediction results for N reflecting elements by executing the same process as the first process for the second idle state / busy state data to the Nth idle state / busy state data instead of the first idle state / busy state data in the first process and all of the (N-1) first machine learning devices.

[0248] Furthermore, the estimation means 43 estimates the achievable rate AR1(t)_ COL ~AR E (t)_ COL are respectively represented as probabilistic neural networks (PNNs). AR _1~PNN AR Input to _E and generate a probabilistic neural network (PNN) AR _1~PNN AR _E executes the flowchart shown in Figure 13 in parallel, and each calculates the achievable rate AR1(t)_ PRD ~AR E (t)_ PRD to the estimation means 43, the estimation means 43 outputs the first achievable rate data to one second machine learning device (PNN AR _1~PNN AR _E), and a first machine learning device PNN calculates an achievable rate corresponding to a maximum probability among probabilities that a first class classified according to the achievable rate of the first achievable rate data matches a second learning stage class that is a class classified according to the achievable rate of the first achievable rate data of the learning stage obtained in the learning stage. AR _1~PNN ARThis corresponds to estimating N achievable rates for N reflecting elements by performing a second process received from one of the (E-1) second machine learning devices (any of the (E-1) second machine learning devices), and performing the same process as the second process on the second achievable rate data through the Nth achievable rate data instead of the first achievable rate data in the second process, and all of the (E-1) second machine learning devices.

[0249] In this case, the AR1(t)_ COL ~AR E (t)_ COL are V reflected beamforming vectors φ1 to φ of the N reflecting elements 21, respectively. V (The reflected beamforming vector φ when one or more reflecting elements 21 are used simultaneously) m ), and constitute "first achievable rate data" through "Eth achievable rate data," each of which arranges the achievable rates in time series.

[0250] In an embodiment of the present invention, the operation of the estimation device 4 may be realized by software. In this case, the estimation device 4 includes a CPU (Central Processing Unit), a ROM (Read Only Memory), and a RAM (Random Access Memory). The ROM stores a program Prog_A consisting of the steps of the flowchart shown in FIG. 11 (including the flowcharts shown in FIGS. 12 and 13).

[0251] The CPU reads out the program Prog_A from the ROM, executes the read out program Prog_A, and selects the usable reflecting elements among the N reflecting elements 21 and the E achievable rates AR1 to AR2. E and the reflected beamforming vector φ of the N reflecting elements 21 is estimated. max The RAM temporarily stores a codebook including the reflection beamforming vectors φ of the N reflection elements 21.

[0252] Furthermore, the program Prog_A may be recorded on a recording medium such as a CD or DVD and distributed. When the recording medium on which the program Prog_A is recorded is attached to a computer, the computer reads and executes the program Prog_A from the recording medium to determine the usable reflecting elements among the N reflecting elements 21 and the E achievable rates AR1 to AR2. E and the reflected beamforming vector φ of the N reflecting elements 21 is estimated. max Get.

[0253] Therefore, the recording medium on which the program Prog_A is recorded is a computer-readable recording medium.

[0254] 14 is a schematic diagram of another wireless communication system according to an embodiment of the present invention, which may be a wireless communication system 10A shown in FIG.

[0255] Referring to FIG. 14, a wireless communication system 10A includes a transmitter 1, IRSs 20 and 30, and a receiver 3.

[0256] The transmitter 1, IRSs 20 and 30, and receiver 3 are arranged in a wireless communication space. The transmitter 1, IRSs 20 and 30, and receiver 3 are arranged in a position where the arrangement [transmitter 1 - IRSs 20 and 30 - receiver 3] is not on a straight line. More specifically, the transmitter 1, IRS 20, and receiver 3 are arranged in a position where the arrangement [transmitter 1 - IRS 20 - receiver 3] is not on a straight line, and the transmitter 1, IRS 30, and receiver 3 are arranged in a position where the arrangement [transmitter 1 - IRS 30 - receiver 3] is not on a straight line.

[0257] Then, when the line connecting the transmitter 1 and the IRS 20 is L3 and the line connecting the IRS 20 and the receiver 3 is L4, the transmitter 1, the IRS 20, and the receiver 3 are arranged so that the angle α1 formed by the lines L3 and L4 is greater than 0. Furthermore, when the line connecting the transmitter 1 and the IRS 30 is L5 and the line connecting the IRS 30 and the receiver 3 is L6, the transmitter 1, the IRS 30, and the receiver 3 are arranged so that the angle α2 formed by the lines L5 and L6 is greater than 0.

[0258] That is, the transmitter 1, the IRS 20, and the receiver 3 are arranged non-linearly in the wireless communication space, and the transmitter 1, the IRS 23, and the receiver 3 are arranged non-linearly in the wireless communication space. As a result, the transmitter 1, the IRS 20, 30, and the receiver 3 are arranged non-linearly in the wireless communication space.

[0259] Each of the IRSs 20 and 30 has the same configuration as the above-mentioned IRS 2. The estimation device 4 of each of the IRSs 20 and 30 receives the received signal y(t) from the receiver 3 and estimates the frequency response FQR_0 of the propagation path from the transmitter 1 to the receiver 3 using the same method as the estimation device 4 of the IRS 2.

[0260] Furthermore, the estimation device 4 of the IRS 20 determines usable reflecting elements among the N reflecting elements 21 of the IRS 20 and E achievable rates AR1 to AR2 according to the flowchart shown in FIG. 11 (including the flowcharts shown in FIGS. 12 and 13). E and the reflected beamforming vector φ of the N reflecting elements 21 of the IRS 20 is estimated. max Get.

[0261] Furthermore, the estimation device 4 of the IRS 30 calculates the usable reflecting elements among the N reflecting elements 21 of the IRS 30 and the E achievable rates AR1 to AR2 in accordance with the flowchart shown in FIG. 11 (including the flowcharts shown in FIGS. 12 and 13). E and the reflected beamforming vector φ of the N reflecting elements 21 of the IRS 30 is estimated. max Get.

[0262] The IRS controller 22 of the IRS 20 is connected to the IRS controller 22 of the IRS 30 by optical fiber, and is capable of mutual communication with the IRS controller 22 of the IRS 30.

[0263] When the communication capacity for wireless communication from the transmitter 1 to the receiver 3 is insufficient using only the N reflecting elements 21 of the IRS 20, the IRS controller 22 of the IRS 20 estimates the amount of resources RSC available for wireless communication using the N reflecting elements 21 of the IRS 20 and the N reflecting elements 21 of the IRS 30.

[0264] Here, the frequency response matrix H(t) of the propagation path estimated by the estimation means 43 of the IRS 20 using equations (1) to (6) is converted into a frequency response matrix H 20 (t), and the frequency response matrix H(t) of the propagation path estimated by the estimation means 43 of the IRS 30 using the equations (1) to (6) is expressed as the frequency response matrix H 30 Let (t).

[0265] The IRS controller 22 of the IRS 20 calculates the frequency response matrix H 30 (t) to the IRS controller 22 of the IRS 30.

[0266] In response to an instruction from the IRS controller 22 of the IRS 20, the IRS controller 22 of the IRS 30 calculates the frequency response matrix H 30 (t) is sent to the IRS controller 22 of the IRS 20.

[0267] The IRS controller 22 of the IRS 20 calculates the frequency response matrix H 30 (t) is received from the IRS controller 22 of the IRS 30.

[0268] Then, in the IRS controller 22 of the IRS 20, the estimation means 43 of the estimation device 4 receives the frequency response matrix H 30 Receive (t).

[0269] The control means 223 of the IRS 20 maintains a codebook CD_B containing the reflected beamforming vectors φ that control the reflection coefficients of the reflecting elements 21 .

[0270] The codebook CD_B has a configuration in which the reflected beamforming vectors φ are arranged in a matrix. The codebook CD_B consists of 2N column elements and, for example, 8 to 16 row elements. Here, the codebook CD_B is assumed to include 8 row elements.

[0271] When the IRS controller 22 of the IRS 20 uses N reflecting elements 21 of the IRS 20 and N reflecting elements 21 of the IRS 30, the IRS controller 22 of the IRS 20 controls 2N reflecting elements 21. The codebook CD_B then includes 8 × 2 reflected beamforming vectors φ.

[0272] The control means 223 of the IRS 20 controls the E when one or more of the 2N reflecting elements 21 are used simultaneously. 2 Reflective element patterns REF_ptn_1 to REF_ptn_E 2 is selected from the codebook CD_B.

[0273] When collecting time series data TSD_1(t)_20, the collection means 41 of the IRS20 receives from the control means 223 of the IRS20 a period ON_period / period OFF_period_n(t) which is a chronological arrangement of a period OFF_period in which one or more of the 2N reflecting elements 21 are turned OFF and a period ON_period in which one or more of the 2N reflecting elements 21 other than the one or more reflecting elements 21 are turned ON.

[0274] Then, the collection means 41 associates "I" with one slot SLT based on the period OFF_period, and associates "B" with one slot SLT based on the period ON_period, and thereafter generates time series data TSD_1(t)_20 in the same manner.

[0275] In this case, the collection means 41 detects that the reflective elements that are turned ON (reflective elements other than one or more reflective elements 21) in one slot SLT are in a busy state B, and detects that the reflective elements that are turned OFF (reflective elements 21) in one slot SLT are in an idle state I, and similarly generates time series data TSD_1(t)_20.

[0276] Then, the collecting means 41 stores the generated time series data TSD_1(t)_20 in the storage means 42. The time series data TSD_1(t)_20 is the time series data B / I1(t) to B / I 2N It consists of (t).

[0277] The collection means 41 of the IRS 20 calculates the achievable rate R using equations (7) to (10). 2 Reflective element patterns REF_ptn_1 to REF_ptn_E 2 and collect the time series data TSD_2(t)_20.

[0278] The time series data TSD_2(t)_20 is AR1(t)_ COL ~AR E^2 (t)_ COL It consists of the following. Note that "E^2" means the square of E.

[0279] FIG. 15 is a schematic diagram of learners 44 and 45 when IRS controller 22 of IRS 20 controls 2N reflecting elements 21.

[0280] Referring to FIG. 15, when the IRS controller 22 of the IRS 20 controls 2N reflecting elements 21, the learner 44 controls 2N probabilistic neural networks PNNs. SLT _1~PNN SLT _2N (see (a) of FIG. 15), and the learning device 45 is 2 Probabilistic neural networks (PNN) AR _1~PNN AR _E 2 (See FIG. 15(b)).

[0281] Probabilistic Neural Network (PNN) SLT _1~PNN SLT Each of the _2N is a probabilistic neural network PNN shown in Figure 6. SLT It has the same structure as _1.

[0282] Probabilistic Neural Network (PNN) SLT _1~PNN SLT _e * (e * is either 1 or 2N) is the time series data B / I e* (t) (time series data B / I1(t)~B / I 2N When any of the inputs (t) is input, the input time series data B / I e* (t) is the prediction matrix P SLT r Class c of the data (time series data TSD_1_learn(t)) stored in (the prediction matrix held in the summation layer) learn The class c when the maximum probability of matching is obtained max (That is, the usability prediction result U e* (t)).

[0283] Therefore, the probabilistic neural network (PNN) SLT _1~PNN SLT _2N are the time series data B / I1(t)~B / I 2N When (t) is input, the class c with the highest probability is obtained. max_1 ~c max_2N (That is, the usability prediction result U1(t)~U 2N (t)) is output (see (a) of Figure 15).

[0284] Also, E 2 Probabilistic neural networks (PNN) AR _1~PNN AR _E 2 Each of these is a probabilistic neural network (PNN) shown in Figure 6. SLT It has the same structure as _1.

[0285] Probabilistic Neural Network (PNN) AR _1~PNN AR _e * (e * is 1~E 2 (either of these) is a time series data AR e* (t)_ COL When input, the prediction matrix P AR r n R n received from rows R The positive integer contained in the row T that received the maximum probability among the probabilities (the positive integer indicating the achievable rate AR) is output.

[0286] Therefore, the probabilistic neural network (PNN) AR _1~PNN AR _E 2 are the time series data AR1(t)_ COL ~AR E^2 (t)_ COL is input, the positive integer contained in the row T that has the greatest probability (i.e., the achievable rate AR1(t)_ PRD ~AR E^2 (t)_ PRD (a positive integer indicating

[0287] In the IRS 20, the estimation means 43 receives from the learning device 44 a usable prediction result indicating usable reflecting elements among the 2N reflecting elements, and 2 Achievable rate AR1(t) of a reflecting element PRD ~AR E^2 (t)_ PRD is received from the learning device 45.

[0288] And in IRS 20, the presumption means 43 is E 2 Achievable rate AR1(t) of a reflecting element PRD ~AR E^2 (t)_ PRD The maximum achievable rate AR max and the maximum achievable rate AR max The reflected beamforming vector φ when maxis detected from the codebook CD_B.

[0289] Then, in the IRS 20, the estimation means 43 estimates the maximum achievable rate AR out of 2N available reflecting elements. max , and the reflected beamforming vector φ max A communication schedule is created using

[0290] In the IRS 20, the estimation means 43 estimates the usable reflecting elements among the 2N reflecting elements, the maximum achievable rate AR max , the reflected beamforming vector φ max And output the communication schedule to the control means 223.

[0291] In the IRS 20, the control means 223 determines the usable reflecting elements among the 2N reflecting elements, the maximum achievable rate AR max , the reflected beamforming vector φ max and a communication schedule from the estimation means 43.

[0292] Then, in the IRS 20, the control means 223 calculates the reflected beamforming vector φ max and the usable reflecting elements among the 2N reflecting elements 21 and the maximum achievable rate AR max The resource amount RSC_20,30 and the communication schedule are transmitted to the transmitter 1.

[0293] The transmitter 1 receives the amount of resources RSC_20,30 and the communication schedule from the IRS controller 22 of the IRS 20.

[0294] When the transmitter 1 starts wireless communication with the receiver 3, the transmitter 1 transmits a communication notification to the IRS controller 22 of the IRS 20.

[0295] The IRS controller 22 of the IRS 20 receives a communication notification from the transmitter 1, and then, when wireless communication starts, calculates a reflected beamforming vector φmax The reflection pattern of the 2N reflecting elements 21 in the IIRS 20, 30 is controlled based on the above.

[0296] In addition, in the embodiment of the present invention, IRS controller 22 of IRS 30 may control 2N reflecting elements in IRSs 20 and 30 in the same manner as IRS controller 22 of IRS 20 .

[0297] The operation of the estimation device 4 of the IRS 20 is performed according to the flowchart shown in Fig. 11 (including the flowcharts shown in Figs. 12 and 13). In this case, "N reflecting elements" should be read as "2N reflecting elements", "N usable reflecting elements" should be read as "2N usable reflecting elements", and "E achievable rates R1 to R2" should be read as "E achievable rates R3 to R4". E " to "E 2 Achievable rates R1 to R E^2 " should be read as "

[0298] Fig. 16 is a schematic diagram showing one specific example of the wireless communication system shown in Fig. 14. Referring to Fig. 16, sidewalks are arranged on both sides of a road.

[0299] IRS-1a, IRS-2a, and IRS-3a are placed facing the sidewalk on one side of the road, and IRS-1a, IRS-2a, and IRS-3a are fixed to the side of the building.

[0300] IRS-1b, IRS-2b, and IRS-3b are placed facing the sidewalk on the other side of the road, and IRS-1b, IRS-2b, and IRS-3b are fixed to the side of the building.

[0301] The IRS controller 1a is connected to the IRS controller 1b and the IRS controller 2a by optical fibers, and also to the IRS-1a and the IRS-2a.

[0302] The IRS controller 2a is connected to the IRS controller 1a and the IRS controller 2b by optical fibers, and is also connected to the IRS-2a and the IRS-3a.

[0303] The IRS controller 1b is connected to the IRS controller 1a and the IRS controller 2b by optical fibers, and also to the IRS-1b and the IRS-2b.

[0304] The IRS controller 2b is connected to the IRS controller 2a and the IRS controller 1b by optical fibers, and is also connected to the IRS-2b and the IRS-3b.

[0305] In FIG. 16, IRS controller 1a, IRS controller 2a, IRS controller 1b, and IRS controller 2b each control N reflective elements 21 of two IRSs.

[0306] The mobile terminal then performs wireless communication using, for example, IRS-2a, IRS-2b, and IRS-3b.

[0307] In this case, the estimation device 4 of the IRS controller 2a estimates the usable reflecting elements among the N reflecting elements 21 of the IRS-2a and the achievable rate AR according to the flowchart shown in FIG. 11 (including the flowcharts shown in FIGS. 12 and 13) before the wireless communication by the mobile terminal is started, and calculates the maximum achievable rate AR in the IRS-2a. max The reflected beamforming vector φ of the N reflecting elements 21 when max Get.

[0308] The estimator 4 of the IRS controller 2a then calculates the maximum achievable rate AR by calculating the number of available reflecting elements among the N reflecting elements 21 of the IRS-2a. max and the reflected beamforming vector φ of the N reflecting elements 21. max Create a communication schedule based on this.

[0309] Then, the estimator 4 of the IRS controller 2a determines the maximum achievable rate AR from the available reflecting elements 21 of the IRS-2a. max and transmitting the communication schedule to the mobile terminal.

[0310] On the other hand, the estimation device 4 of the IRS controller 2b determines the usable reflecting elements among the N reflecting elements 21 of the IRS-2b and the maximum achievable rate AR according to the flowchart shown in FIG. 11 (including the flowcharts shown in FIGS. 12 and 13) before the mobile terminal starts wireless communication. max and the reflected beamforming vector φ of the N reflecting elements 21 of the IRS-2b is calculated. max Get.

[0311] Furthermore, before wireless communication by the mobile terminal is initiated, the estimation device 4 of the IRS controller 2b determines which of the N reflecting elements 21 of the IRS-3b are available and which are the maximum achievable rate AR according to the flowchart shown in FIG. 11 (including the flowcharts shown in FIGS. 12 and 13). max and the reflected beamforming vector φ of the N reflecting elements 21 of the IRS-3b is calculated. max Get.

[0312] The estimator 4 of the IRS controller 2b then calculates the maximum achievable rate AR by using the available reflecting elements among the N reflecting elements 21 of the IRS-2b. max and the reflected beamforming vector φ of the N reflecting elements 21. max and the usable reflecting elements among the N reflecting elements 21 of the IRS-3b, and the maximum achievable rate AR max and the reflected beamforming vector φ of the N reflecting elements 21. max A communication schedule is created based on the above.

[0313] Then, the estimator 4 of the IRS controller 2b determines the usable reflecting elements among the N reflecting elements 21 of the IRS-2b and the maximum achievable rate AR maxand the usable reflecting elements among the N reflecting elements 21 of the IRS-3b and the maximum achievable rate AR max and the communication schedule to the mobile terminal.

[0314] Then, the IRS controller 2a and the IRS controller 2b receive a communication notification from the mobile terminal. Then, when wireless communication by the mobile terminal is started, the IRS controller 2a calculates the reflected beamforming vector φ of the IRS-2a. max The IRS controller 2b controls the reflection pattern of the N reflecting elements of the IRS-2a according to the following equation: max The reflection patterns of the N reflecting elements of IRS-2b are controlled according to the following equation, and the reflection beamforming vector φ of IRS-3b is also controlled according to the following equation: max The reflection pattern of the N reflection elements of the IRS-3b is controlled according to the above.

[0315] This allows a mobile terminal to use three IRS-2a, IRS-2b, and IRS-3b in a wireless communication space where buildings are located on both sides of a road to conduct wireless communication while preventing the radio wave's reach from being shortened by the buildings.

[0316] In one example of a wireless communication system shown in FIG. 16, the estimator 4 of the IRS controller 2a calculates the usable reflecting elements among the N reflecting elements of the IRS-2a and the maximum achievable rate AR of the N reflecting elements. max The estimator 4 of the IRS controller 2b estimates the usable reflecting elements among the N reflecting elements of the IRS-2b and the maximum achievable rate AR of the N reflecting elements. max The maximum achievable rate AR of the N reflecting elements of the IRS-3b is estimated. max If we estimate the number of reflecting elements that can be used among the estimated N reflecting elements of the IRS-2a and the maximum achievable rate AR of the N reflecting elements, max and the estimated usable reflecting elements among the N reflecting elements of the IRS-2b and the maximum achievable rate AR of the N reflecting elements.max and the estimated usable reflecting elements among the N reflecting elements of the IRS-3b and the maximum achievable rate AR of the N reflecting elements. max Based on this, communication resources can be secured using the N reflecting elements of IRS-2a, the N reflecting elements of IRS-2b, and the N reflecting elements of IRS-3b, so that a schedule for using the reflecting elements can be created by determining in advance whether enough reflecting elements are actually available to ensure a sufficient transmission rate.

[0317] According to the above-described embodiment, the estimation device according to the embodiment of the present invention comprises: M T (M T a transmitter having N (N is an integer of 2 or more) antennas, a first IRS which is an electromagnetic wave reflector having N (N is an integer of 2 or more) reflecting elements whose reflection characteristics can be dynamically changed, and M R (M R An estimation device for estimating information useful for predicting a usage schedule, which is a schedule for using N reflecting elements, in a wireless communication system in which a receiver having N (an integer equal to or greater than 2) antennas is non-linearly arranged, the estimation device comprising: A first time series data showing time series data for each of N reflective elements, the time series data being a busy state in which the reflective element is in use and an idle state in which the reflective element is not in use; T and second time series data indicative of an achievable rate, the achievable transmission rate of the N reflecting elements when transmitting a signal using the N antennas; The system may include an estimation means that performs a first estimation process that uses a first learning device to estimate a usability prediction result for each of N reflecting elements, which is the result of predicting whether or not the reflecting elements in the first IRS are usable, based on first time series data, and a second estimation process that uses a second learning device to estimate the achievable rates of the N reflecting elements of the first IRS in the propagation path for reflecting elements that are in use among the N reflecting elements, based on second time series data.

[0318] If the estimation means estimates the predicted results of availability of N of the N reflecting elements and the achievable rate of the reflecting elements that are being used among the N reflecting elements, resources for communication using the N reflecting elements can be secured based on the estimated predicted results of availability of N and the achievable rate, so that a schedule for using the reflecting elements can be created by determining in advance whether enough reflecting elements are actually available to ensure a sufficient transmission rate.

[0319] Furthermore, the program according to the embodiment of the present invention is M T (M T a transmitter having N (N is an integer of 2 or more) antennas, a first IRS which is an electromagnetic wave reflector having N (N is an integer of 2 or more) reflecting elements whose reflection characteristics can be dynamically changed, and M R (M R A program for causing a computer to execute estimation of information useful for predicting a usage schedule, which is a schedule for using N reflecting elements, in a wireless communication system in which a receiver having N (an integer greater than or equal to 2) antennas is non-linearly arranged, the program comprising: The collecting means collects first time series data indicating time series data for each of the N reflecting elements, the time series data being composed of a busy state in which the reflecting element is in use and an idle state in which the reflecting element is not in use, and T a second time series of data indicative of an achievable transmission rate of the N reflecting elements when transmitting a signal using the N antennas; The estimation means may cause the computer to execute a second step of performing a first estimation process in which the estimation means uses a first learning device to estimate a usability prediction result for each of N reflecting elements, which is a result of predicting whether or not the reflecting elements in the first IRS are usable, based on first time series data, and a second estimation process in which the estimation means uses a second learning device to estimate the achievable rates of the N reflecting elements of the first IRS in the propagation path for those reflecting elements that are in use among the N reflecting elements, based on second time series data.

[0320] If the program causes a computer to estimate the predicted results of N usability of N reflecting elements and the achievable rate for the reflecting elements that are being used among the N reflecting elements, resources for communication using the N reflecting elements can be secured based on the estimated predicted results of N usability and the achievable rate for the reflecting elements that are being used among the N reflecting elements. This makes it possible to create a schedule for using the reflecting elements by determining in advance whether enough reflecting elements are actually available to ensure a sufficient transmission rate.

[0321] In the embodiment of the present invention, N usability prediction results of N reflecting elements 21 and achievable rates R1 to R2 of the reflecting elements being used among the N reflecting elements are calculated. N constitutes "information useful for predicting a usage schedule, which is a schedule for using N reflecting elements 21." Note that the reflected beam forming vectors φ max may be included in "information useful for predicting a usage schedule, which is a schedule for using N reflecting elements 21."

[0322] In addition, in this embodiment of the present invention, learning device 44 constitutes a "first learning device," and learning device 45 constitutes a "second learning device."

[0323] Furthermore, in the embodiment of the present invention, a probabilistic neural network (PNN) SLT _1~PNN SLT _N constitutes the "N first machine learning machines" and is a probabilistic neural network (PNN) AR _1~PNN AR _E constitutes "E second machine learning machines."

[0324] Furthermore, in this embodiment of the present invention, the time series data B / I1(t) constitutes "first idle state / busy state data", and the time series data B / I N (t) constitutes the "Nth idle / busy state data."

[0325] The embodiments disclosed herein should be considered to be illustrative in all respects and not restrictive. The scope of the present invention is defined by the claims, not by the description of the above embodiments, and is intended to include all modifications within the meaning and scope of the claims. [Industrial Applicability]

[0326] The present invention is applied to an estimation device, a wireless communication system including the same, and a program to be executed by a computer. [Explanation of symbols]

[0327] 1 transmitter, 2, 20, 30 IRS, 3 receiver, 4 estimator, 10, 10A wireless communication system, 21 reflecting element, 22 IRS controller, 41 collecting means, 42 storing means, 43 estimating means, 44, 45 learning device, 221 antenna, 222 communication means, 223 control means.

Claims

1. M T (M T a transmitter having N (N is an integer of 2 or more) antennas, a first IRS which is an electromagnetic wave reflector having N (N is an integer of 2 or more) reflecting elements whose reflection characteristics can be dynamically changed, and M R (M R an estimation device for estimating information useful for predicting a usage schedule, which is a schedule for using N reflecting elements, in a wireless communication system in which a receiver having N antennas (where N is an integer greater than or equal to 2) is non-linearly arranged; a first time series data indicating, for each of the N reflecting elements, time series data consisting of a busy state in which the reflecting element is in use and an idle state in which the reflecting element is not in use; T and second time series data indicative of an achievable rate, the achievable transmission rate of the N reflecting elements when transmitting a signal using the N antennas; An estimation device comprising an estimation means that performs a first estimation process that uses a first learning device to estimate a usability prediction result for each of the N reflecting elements, which is a result of predicting whether the reflecting elements in the first IRS are usable or not, based on the first time series data, and a second estimation process that uses a second learning device to estimate the achievable rate of the N reflecting elements of the first IRS for used reflecting elements among the N reflecting elements in a propagation path in which the transmitter, the first IRS, and the receiver are arranged non-linearly, based on the second time series data.

2. 2. The estimation device according to claim 1, wherein, in the first estimation process, the estimation means estimates usable reflecting elements among the N reflecting elements by inputting the first time series data to the first learning device and receiving from the first learning device the idle state corresponding to the maximum probability among probabilities that a first class classified according to the idle state of the first time series data matches a first learning stage class that is a class classified according to the idle state of first learning time series data acquired in a learning stage; and in the second estimation process, the estimation means estimates the achievable rate of a used reflecting element among the N reflecting elements by inputting the second time series data to the second learning device and receiving from the second learning device the achievable rate corresponding to the maximum probability among probabilities that a second class classified according to the achievable rate of the second time series data matches a second learning stage class that is a class classified according to the achievable rate of second learning time series data acquired in a learning stage.

3. the first learning device comprises N first machine learning devices, the second learning device comprises E second machine learning devices, the number of which is the number of reflection patterns of one or more of the N reflection elements when the one or more reflection elements are simultaneously used, the first time series data comprises first idle state / busy state data to Nth idle state / busy state data, each of which arranges the idle state and the busy state in time series, the second time series data corresponds to E reflection patterns and comprises first achievable rate data to Eth achievable rate data, each of which arranges the achievable rates in time series, The estimation means In the first estimation process, the first idle state / busy state data is input to one first machine learning machine among the N first machine learning machines, and a first process is performed to receive from the one first machine learning machine the idle state corresponding to the maximum probability among probabilities that the first class classified according to the idle state of the first idle state / busy state data matches the first learning stage class, which is the class classified according to the idle state of the first idle state / busy state data of the learning stage acquired in a learning stage, and the same process as the first process is performed for second idle state / busy state data to Nth idle state / busy state data among the first idle state / busy state data to Nth idle state / busy state data instead of the first idle state / busy state data in the first process, and for all of the (N-1) first machine learning machines, thereby estimating the N usability prediction results of the N reflecting elements; 3. The estimation device according to claim 2, wherein, in the second estimation process, the first achievable rate data is input to one of the second machine learning machines, and a second process is performed to receive from the one first machine learning machine the achievable rate corresponding to the largest probability among probabilities that the first class classified according to the achievable rate of the first achievable rate data matches the second learning stage class, which is a class classified according to the achievable rate of the first achievable rate data of the learning stage acquired in a learning stage, and the achievable rate of a used reflecting element among the N reflecting elements is estimated by performing the same process as the second process for second achievable rate data through E achievable rate data among the first achievable rate data through E achievable rate data and all of the (E−1) second machine learning machines, instead of the first achievable rate data in the second process.

4. the N first machine learning machines execute the first process in parallel in the first estimation process; The estimation device according to claim 3 , wherein the E second machine learning machines execute the second process in parallel in the second estimation process.

5. each of the N first machine learning machines classifies the idle states acquired in a first learning stage into a first plurality of classes and stores the classes in a first prediction matrix, and when one of the idle state / busy state data is input, selects one of the classes from the first plurality of classes stored in the first prediction matrix that has the highest probability of matching the idle state of the input one of the idle state / busy state data, and outputs the availability prediction result to the estimation means based on the selection of the one of the classes; 5. The estimation device according to claim 3, wherein each of the E second machine learning machines classifies the achievable rates obtained in the second learning stage into a second plurality of classes and stores them in a second prediction matrix, and when one piece of achievable rate data is input, selects the one class from the second plurality of classes that has the highest probability of matching the achievable rate of the input piece of achievable rate data, and outputs the achievable rate classified into the one selected class to the estimation means.

6. In the second estimation process, the estimation means R The propagation paths to the antennas are divided into M R Transformed into the independent M propagation paths R M propagation paths R Calculate the transmission rates of M R The estimation device according to claim 1 , wherein the achievable rate is a sum of transmission rates.

7. In the second estimation process, the estimation means R Based on the signal-to-interference-plus-noise power ratios of the antennas, R The estimation device according to claim 6 , which calculates transmission rates.

8. A radio communication system comprising: a first estimating device comprising the estimating device according to any one of claims 1 to 7; the transmitter; the receiver; and the first IRS; The first estimator is located within the first IRS.

9. A second estimation device comprising the estimation device according to any one of claims 1 to 7, and an N * (N * a second IRS that is an electromagnetic wave reflector having a number of reflection elements (where the number is an integer greater than or equal to 2); The wireless communication system of claim 8 , wherein the second estimator is located within the second IRS.

10. M T (M T a transmitter having N (N is an integer of 2 or more) antennas, a first IRS which is an electromagnetic wave reflector having N (N is an integer of 2 or more) reflecting elements whose reflection characteristics can be dynamically changed, and M R (M R a program for causing a computer to execute estimation of information useful for predicting a usage schedule, which is a schedule for using N reflecting elements, in a wireless communication system in which a receiver having N (an integer equal to or greater than 2) antennas is non-linearly arranged; a first time series data indicating, for each of the N reflecting elements, time series data consisting of a busy state in which the reflecting element is in use and an idle state in which the reflecting element is not in use; T a first step in which a second time series of data indicative of an achievable rate is collected, the achievable transmission rate of the N reflecting elements when transmitting a signal using the N antennas; a first estimation process in which a usability prediction result is estimated for each of the N reflecting elements, the usability prediction result being a result of predicting whether the reflecting element in the first IRS is usable or not, based on the first time series data, using a first learning device; and a second estimation process in which a usability prediction result is estimated for each of the N reflecting elements, the usability prediction result being a result of predicting whether the reflecting element in the first IRS is usable or not, based on the second time series data, using a second learning device. A program for causing a computer to execute a second estimation process in which the achievable rate of the N reflecting elements of an IRS is estimated for a reflecting element that is being used among the N reflecting elements.

11. A program for being executed by a computer as described in claim 10, wherein in the first estimation process of the second step, the first time series data is input to the first learning device, and usable reflecting elements among the N reflecting elements are estimated by outputting from the first learning device the idle state corresponding to the maximum probability among probabilities that a first class classified according to the idle state of the first time series data matches a first learning stage class, which is a class classified according to the idle state of the first learning time series data obtained in the learning stage; and in the second estimation process of the second step, the second time series data is input to the second learning device, and the achievable rate corresponding to the maximum probability among probabilities that a second class classified according to the achievable rate of the second time series data matches a second learning stage class, which is a class classified according to the achievable rate of the second learning time series data obtained in the learning stage, thereby estimating the achievable rate of a used reflecting element among the N reflecting elements.

12. the first learning device comprises N first machine learning devices, the second learning device comprises E second machine learning devices, the number of which is the number of reflection patterns of one or more of the N reflection elements when the one or more reflection elements are simultaneously used, the first time series data comprises first idle state / busy state data to Nth idle state / busy state data, each of which arranges the idle state and the busy state in time series, the second time series data corresponds to E reflection patterns and comprises first achievable rate data to Eth achievable rate data, each of which arranges the achievable rates in time series, In the first estimation process of the second step, the first idle state / busy state data is input to one first machine learning machine among N first machine learning machines, and a first process is executed in which the idle state corresponding to the maximum probability among probabilities that the first class classified according to the idle state of the first idle state / busy state data matches the first learning stage class, which is the class classified according to the idle state of the first idle state / busy state data of the learning stage acquired in a learning stage, is output from the one first machine learning machine, and the same process as the first process is executed for second idle state / busy state data to Nth idle state / busy state data among the first idle state / busy state data to Nth idle state / busy state data instead of the first idle state / busy state data in the first process, and for all of the (N-1) first machine learning machines, thereby estimating the N usability prediction results of the N reflecting elements, 12. The program for causing a computer to execute the program according to claim 11, wherein in the second estimation process of the second step, the first achievable rate data is input to one of the second machine learning machines, and a second process is executed in which the achievable rate corresponding to the largest probability of the first class classified according to the achievable rate of the first achievable rate data matching the second learning stage class, which is a class classified according to the achievable rate of the first achievable rate data of the learning stage acquired in a learning stage, is output from the one first machine learning machine, and ... of the N reflective elements used is estimated by executing the same process as the second process for second achievable rate data to E achievable rate data among the first achievable rate data to E achievable rate data and all of the (E-1) second machine learning machines, instead of the first achievable rate data in the second process.

13. In the first estimation process of the second step, the first process is executed in parallel, The program for causing a computer to execute the program according to claim 12 , wherein the second estimation process in the second step is executed in parallel with the second process.

14. The idle states acquired in the first learning stage are classified into a first plurality of classes and stored in a first prediction matrix, and when one of the idle state / busy state data is input, one of the classes having the highest probability of matching the idle state of the input one of the idle state / busy state data is selected from the first plurality of classes stored in the first prediction matrix, and the usability prediction result is output based on the selection of the one of the class; 14. A program for causing a computer to execute the program described in claim 12 or 13, wherein the achievable rates obtained in the second learning stage are classified into a second plurality of classes and stored in a second prediction matrix, and when one of the achievable rate data is input, the one of the classes that has the highest probability of matching the achievable rate of the input one of the achievable rate data is selected from the second plurality of classes, and the achievable rate classified into the selected one of the classes is output.

15. In the second estimation process of the second step, the M of the receiver R The propagation paths to the antennas are M independent. R are converted into the independent M propagation paths R M propagation paths R The transmission rates M R 15. The program for causing a computer to execute the program according to claim 10, wherein the achievable rate is a sum of the transmission rates.

16. In the second estimation process of the second step, R Based on the signal-to-interference-plus-noise power ratios of the antennas, R The program executed by a computer according to claim 15, wherein the transmission rates are calculated.

Images