Electronic device and method of operating electronic device

Abstract

An electronic device and a method of operating the electronic device are provided. The electronic device includes: a memory; a communication module comprising an antenna array configured to form a plurality of candidate beams; and a processor connected to the communication module and the memory, the processor comprising a neural network-based channel state classifier, where the processor is configured to: generate channel characteristic data based on information measured from the plurality of candidate beams for a signal received from a base station, channel characteristic data is input to a channel state classifier to infer a channel state between the electronic device and a base station, and the plurality of candidate beams are reformed based on the inferred channel state.

Description

[0001] This application is based on and claims the benefit of priority from Korean Patent Application No. 10-2023-0097060 filed on July 25, 2023, in the Korean Intellectual Property Office, and Korean Patent Application No. 10-2023-0161432 filed on November 20, 2023, in the Korean Intellectual Property Office, the disclosures of which are incorporated herein in their entirety by reference. Technical Field

[0002] The disclosure relates to an electronic device for performing beam management according to a channel state based on a neural network in a wireless communication system and a method of operating the electronic device. Background Art

[0003] In a wireless communication system, an electronic device (e.g., a terminal) may send data to and receive data from an external device (e.g., a base station) via a beam formed by concentrating a signal in a specific direction (e.g., beamforming). The electronic device may select an optimal beam to be applied to an antenna array for wireless communication through a beam training process, which predetermines a candidate beam and compares the measurement results of the signal quality between the beam of the external device and the candidate beam of the electronic device. Currently, when a beam with a small beam width (hereinafter referred to as a narrow beam) is used as a candidate beam, the electronic device may increase the beam matching gain between the external device and the electronic device. However, when a narrow beam is used as at least one of the candidate beams of the electronic device, the number of candidate beams whose signal quality needs to be measured during beam training may increase. Therefore, for all of the increased number of candidate beams, excessive overhead may occur during the beam training process. Such excessive overhead may lead to problems that degrade the performance of the entire electronic device or degrade the communication quality between the electronic device and the external device. Summary of the invention

[0004] Embodiments of the present disclosure provide an electronic device and a method of operating the electronic device for applying channel characteristic data measured from a plurality of candidate beams of the electronic device to a pre-trained neural network to infer a channel state and perform beam management corresponding to the inferred channel state.

[0005] Problems to be solved by the technical idea of ​​the embodiments of the present disclosure are not limited to the above-mentioned problems, and other problems not mentioned will be clearly understood by those skilled in the art from the following description.

[0006] According to one aspect of the disclosure, an electronic device includes: a memory; a communication module including an antenna array configured to form multiple candidate beams; and a processor connected to the communication module and the memory, the processor including a neural network-based channel state classifier, wherein the processor is configured to: generate channel characteristic data based on information measured from the multiple candidate beams for signals received from a base station, input the channel characteristic data into the channel state classifier to infer the channel state between the electronic device and the base station, and reform the multiple candidate beams based on the inferred channel state.

[0007] According to one aspect of the disclosure, a method for operating an electronic device includes: generating channel characteristic data based on information measured from multiple candidate beams for a signal received from a base station; applying the channel characteristic data to a neural network based on a deep learning algorithm to infer a channel state between the electronic device and the base station; and reforming the multiple candidate beams based on the inferred channel state.

[0008] According to one aspect of the disclosure, a method for operating an electronic device includes: forming a plurality of candidate beams for performing wireless communication with an external device; generating first channel characteristic data representing channel characteristics in a spatial domain and second channel characteristic data representing channel characteristics in a time domain based on information measured from the plurality of candidate beams; applying data obtained by combining the first channel characteristic data and the second channel characteristic data in a two-dimensional form to a pre-trained channel state inference model to infer a channel state between the electronic device and the external device; and reforming the plurality of candidate beams based on the channel state, wherein the first channel characteristic data includes information about a reference signal received power (RSRP) measured from the plurality of candidate beams, and wherein the second channel characteristic data includes information about one of a power delay profile (PDP) and a channel impulse response (CIR) measured from the plurality of candidate beams. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] The above and other aspects, features and advantages of specific embodiments of the present disclosure will become more apparent from the following description in conjunction with the accompanying drawings.

[0010] Figure 1 An example of channel status in a wireless communication environment according to one or more embodiments is shown.

[0011] Figure 2 is a block diagram schematically illustrating the operation of an electronic device according to one or more embodiments.

[0012] Figure 3 is a block diagram illustrating an electronic device according to one or more embodiments.

[0013] Figure 4is a flowchart illustrating a method of operating an electronic device according to one or more embodiments.

[0014] Figure 5A An example of a channel state inference operation of an electronic device according to one or more embodiments is shown.

[0015] Figure 5B Examples of channel characteristic data according to one or more embodiments are shown.

[0016] Figure 5C An example of output data of a channel state classifier according to one or more embodiments is shown.

[0017] Fig. 6A An example of a channel state inference operation of an electronic device according to one or more embodiments is shown.

[0018] Figure 6B Examples of channel characteristic data according to one or more embodiments are shown.

[0019] Figure 7 is a flowchart illustrating a method of operating an electronic device according to one or more embodiments.

[0020] Figure 8 is a flowchart illustrating a method of operating an electronic device according to one or more embodiments.

[0021] Fig. 9 is a flowchart illustrating a method of operating an electronic device according to one or more embodiments.

[0022] Fig.10 is a flowchart illustrating a method of operating an electronic device according to one or more embodiments.

[0023] Fig.11 is a flowchart illustrating a method of operating an electronic device according to one or more embodiments.

[0024] Fig.12 is a flowchart illustrating a method of operating an electronic device according to one or more embodiments.

[0025] Fig.13A is a diagram illustrating a beam strategy corresponding to a channel state inferred by an electronic device according to one or more embodiments.

[0026] Fig. 13B Examples of candidate beams reformed by an electronic device are shown according to one or more embodiments.

[0027] Fig.14A An example of a beam management operation based on a channel status of an electronic device according to one or more embodiments is shown.

[0028] Fig. 14B An example of a beam training operation using candidate beams reformed for each channel state according to one or more embodiments is shown.

[0029] Fig. 14C An example of a beam training operation using candidate beams reformed for each channel state according to one or more embodiments is shown.

[0030] Fig.15 is a schematic block diagram of a wireless communication system according to one or more embodiments. DETAILED DESCRIPTION

[0031] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings.

[0032] For convenience, according to an embodiment of the present disclosure, some of the terms and names defined in the 3rd Generation Partnership Project (3GPP) Long Term Evolution (LTE) standard or the New Radio (NR, or referred to as the New Air Interface) standard are used. However, the embodiments of the present disclosure are not limited to the terms and names, and can be equally applied to systems according to other standards known to those of ordinary skill in the art.

[0033] As a non-limiting example, a wireless communication system according to one or more embodiments may include an NR system, a fifth generation (5G) system, a long term evolution (LTE) system, a code division multiple access (CDMA) system, a global system for mobile communications (GSM) system, a wireless local area network (WLAN) system, or any other wireless communication system. Hereinafter, the wireless communication system is described on the premise that the wireless communication system is an NR system, an LTE system, or a system configured to support communications based on NR and LTE. However, embodiments of the present disclosure are not limited thereto.

[0034] An electronic device (eg, terminal) according to one or more embodiments may be referred to as, for example, user equipment, a mobile station (MS), a mobile terminal (MT), a user terminal, a subscriber station (SS), a wireless device, or a handheld device.

[0035] The electronic device according to one or more embodiments may support 4G communication (eg, LTE and LTE-Advanced (LTE-A)) or 5G communication (eg, NR) defined in the 3GPP standard.

[0036] In one or more embodiments, for 4G communication and 5G communication, the electronic device may support a CDMA-based communication protocol, a wideband CDMA (WCDMA)-based communication protocol, a time division multiple access (TDMA)-based communication protocol, a frequency division multiple access (FDMA)-based communication protocol, an orthogonal frequency division multiplexing (OFDM)-based communication protocol, a cyclic prefix (CP)-OFDM-based communication protocol, a discrete Fourier transform-spread-OFDM-based communication protocol, a non-orthogonal multiple access (NOMA)-based communication protocol, and a generalized frequency division multiplexing (GFDM)-based communication protocol.

[0037] In one or more embodiments, wireless communication may be performed based on one of 5G communication (e.g., FR1 band or FR2 band of 5G communication), 4G communication (LTE or LTE-A), ultra-wideband (UWB), and wireless fidelity (Wi-Fi) that is capable of performing high-speed data transmission between an external device (e.g., a base station) and an electronic device.

[0038] According to one or more embodiments, an external device may represent a fixed station that communicates with an electronic device and / or another base station. For example, the external device may be referred to as a Node B, an evolved Node B (eNB), a sector, a site, a base transceiver system (BTS), an access point (AP), a relay node, a remote radio head (RRH), or a radio unit (RU). For example, in a wireless communication system, an external device as a minimum unit of an area for providing communication services to each external device includes at least one cell. The external device may provide efficient multiple access communications for multiple terminals based on specific frequency resources allocated to each cell. For example, in the case of a multiple-input multiple-output (MIMO) system, the electronic device and the external device may perform wireless communications based on multiple component carriers.

[0039] Here, the reference signal received power (RSRP) pattern data may represent the data including the power received from the electronic device 100 (reference signal received power) for the signal received from the external device. Figure 1 ) data of RSRP values ​​measured for multiple candidate beams.

[0040] Here, the neural network may refer to a neural network that is pre-trained to receive channel characteristic data measured from a plurality of candidate beams of the electronic device 100 (e.g., one of RSRP pattern data, {RSRP pattern data and power delay profile (PDP) data}, and {RSRP pattern data and channel impulse response (CIR) data}) to infer a channel state corresponding to the channel characteristic data through a deep learning algorithm.

[0041] Here, the sensor-assisted beam management (BM) function may refer to a function of generating compensation information according to a change in the position / orientation (or referred to as orientation) of the electronic device 100 and correcting an existing target beam based on the generated compensation information.

[0042] Figure 1 An example of a wireless communication environment is shown in accordance with one or more embodiments.

[0043] Reference Figure 1 In a wireless communication environment according to one or more embodiments, the electronic device 100 may transmit and receive signals to and from the base station 200 through a channel of a first path 101 having a line-of-sight (LoS) propagation characteristic. For example, the channel state of the first path 101 may include a LoS state. For example, a signal received through a channel of the first path 101 may be directly received by the electronic device 100 after being transmitted from the base station 200 without being reflected by another object (e.g., the first object 310 or the second object 320). A signal received through a channel of the first path 101 may be received by the electronic device 100 through a channel in which it is assumed that physical phenomena such as diffraction, scattering, and reflection do not occur in a propagation path.

[0044] The electronic device 100 may transmit and receive signals to and from the base station 200 through the channels of the second path 102 and the third path 103 having non-line-of-sight (NLoS) propagation characteristics. For example, the channel states of the second path 102 and the third path 103 may include an NLoS state. For example, a signal received through the channels of the second path 102 and the third path 103 may be transmitted by the base station 200, then reflected by the first object 310 and received by the electronic device 100 (e.g., the second path 102), or may be transmitted by the base station 200, then reflected by the second object 320 and received by the electronic device 100 (e.g., the third path 103). When physical phenomena such as diffraction, scattering, and reflection occur in the propagation path, the signal received through the channels of the second path 102 and the third path 103 may be received by the electronic device 100 through two or more channels.

[0045] The electronic device and method according to one or more embodiments may infer the channel state (or channel environment) of the channel receiving the signal by using a pre-trained neural network, and may perform adaptive beam management based on the inferred channel state. Figures 2 to 15 As described above, the apparatus and method according to one or more embodiments can prevent excessive overhead in the beam training process by performing adaptive beam management on the channel status, and can improve the communication quality by applying a beam strategy corresponding to the channel status.

[0046] Figure 2 is a block diagram schematically illustrating the operation of an electronic device according to one or more embodiments.

[0047] Reference Figure 2 , the electronic device 100 may include a processor 110 , a channel state classifier 111 including a neural network, and a communication module 120 . Figure 1 Some components of the electronic device 100 are shown in FIG. However, the embodiments of the present disclosure are not limited thereto, and the electronic device 100 may further include other components (see FIG. Figure 3 and Fig.15 ).

[0048] exist Figure 2 In the present invention, the communication module 120 of the electronic device 100 may form a predetermined number of candidate beams (hereinafter referred to as multiple candidate beams) according to the characteristics of the wireless communication system or the wireless communication situation. For example, based on the wireless communication system or the wireless communication situation, the candidate beams of the electronic device 100 may include all beams in a predefined beam codebook or some beams in a predefined beam codebook. The multiple candidate beams of the electronic device 100 may be arranged to be non-overlapping to cover a spatial area in which a signal can be received from an external device based on a predetermined number of candidate beams.

[0049] The communication module 120 may generate channel characteristic data 10 including information measured from a plurality of candidate beams, and may transmit the channel characteristic data 10 to the channel state classifier 111. In one or more examples, the channel characteristic data 10 may include information on at least one of RSRP, power delay profile (PDP), and channel impulse response (CIR) measured from a plurality of candidate beams for a signal received from an external device.

[0050] In one or more examples, the channel state classifier 111 may infer the channel state 20 by applying the channel characteristic data 10 received from the communication module 120 to a neural network. The neural network may be pre-trained to receive the channel characteristic data (e.g., RSRP pattern data, {RSRP pattern data and PDP data}, or {RSRP pattern data and CIR data}) 10 and infer the channel state 20 corresponding to the channel characteristic data 10 through a deep learning algorithm. The channel state classifier 111 may select a method for inferring the channel state 20 based on whether channel information about the channel characteristic data 10 (e.g., a label value corresponding to the channel characteristic data 10) is obtained. According to one or more embodiments, the following reference is made to Figure 7 A detailed description of the operation of the channel state classifier 111 selecting an inference method is described.

[0051] In one or more examples, the processor 110 may perform beam management for updating a candidate beam set (including a plurality of candidate beams) of the electronic device 100 based on the channel state 20 output from the channel state classifier 111. For example, as described below with reference to Figures 8 to 13B The processor 110 may be based on the beam strategy corresponding to the channel state 20 inferred by the channel state classifier 111 (refer to Fig.13A ) Re-form multiple candidate beams by controlling the communication module 120.

[0052] Devices and methods according to various embodiments may provide an electronic device for minimizing beam training overhead and improving beam performance gain through adaptive beam management according to channel status through the following operation: channel status of channel characteristic data measured from multiple candidate beams of the electronic device is inferred by using a neural network trained through a deep learning algorithm.

[0053] Figure 3 is a block diagram illustrating an electronic device 100 according to one or more embodiments.

[0054] Reference Figure 3 , the electronic device 100 may include a processor 110 , a communication module 120 , a sensor 130 , and a memory 140 , wherein the processor 110 includes a channel state classifier 111 .

[0055] In one or more embodiments, the processor 110 may execute a program and / or process stored in the memory 140 to perform control operations of the electronic device 100. For example, a program executed to perform operations of the neural network 112 training the channel state classifier 111 according to one or more embodiments, an operation of inferring a channel state based on the trained neural network 112, and performing beam management (e.g., updating beam strategies for multiple candidate beams) based on the inferred channel state may be stored in the memory 140 as instructions or program codes executed by the processor 110. The processor 110 may access the memory 140 and execute the stored program code to control the electronic device 100 to infer a channel state corresponding to channel characteristic data measured from multiple candidate beams of the communication module 120 through the neural network 112, and perform beam management (e.g., updating beam strategies for multiple candidate beams) based on the inferred channel state.

[0056] In one or more examples, the processor 110 may include a channel state classifier 111. The channel state classifier 111 may be implemented by a logic block implemented by logic synthesis, a software block executed by a processor, or a combination thereof. The channel state classifier 111 may include a program as a set of multiple instructions executed by the processor 110, and the program may be stored in a memory accessible by the processor 110.

[0057] In one or more examples, the channel state classifier 111 may be pre-trained to apply channel characteristic data measured from a plurality of candidate beams to the neural network 112 to infer a channel state corresponding to the channel characteristic data. The neural network 112 may be trained to infer the channel state through a deep learning algorithm. The neural network 112 may have a multi-layer perceptron (MLP), a convolutional neural network (CNN), a recurrent neural network (RNN), or a transformer structure.

[0058] In one or more examples, the channel state classifier 111 may infer a channel state corresponding to the channel characteristic data measured from the plurality of candidate beams for a signal received from an external device by using the neural network 112. In one or more examples, the channel state as an output of the channel state classifier 111 may include a channel state according to various communication environments / scenarios and a LoS state / NLoS state (see Fig.13A ), refer to the following Figure 5A and Fig. 6A Example embodiments thereof are described.

[0059] In one or more examples, the processor 110 may perform optimal beam management corresponding to the channel state based on the channel state output from the channel state classifier 111. For example, the processor 110 may control the communication module 120 to update the candidate beam set of the electronic device 100 using the beam strategy corresponding to the channel state. The processor 110 may reform a plurality of candidate beams included in the candidate beam set of the electronic device 100 based on the beam strategy corresponding to the inferred channel state, as described below with reference to Figures 7 to 12 Example embodiments thereof are described.

[0060] In one or more examples, the processor 110 may perform beam training by using the updated plurality of candidate beams, may select a target beam from the updated plurality of candidate beams based on the beam training result, and may perform wireless communication with an external device through the target beam.

[0061] When the position or orientation of the electronic device 100 is changed after the target beam is selected, the processor 110 may generate compensation information about the target beam based on the sensing data received from the sensor 130 (e.g., data about the amount of change in the orientation of the electronic device 100), and may correct the existing target beam based on the generated compensation information. The processor 110 may select the corrected target beam as the final target beam, and may perform wireless communication with an external device. In the current specification, as described above, the function of generating compensation information according to the change in the position / orientation of the electronic device 100 to correct the existing target beam is defined as a sensor-assisted beam management function.

[0062] In one or more examples, the communication module 120 includes at least one antenna array including a plurality of antennas, and a plurality of candidate beams may be formed by using a plurality of antennas (or antenna arrays). The communication module 120 may send a signal to an external device through a plurality of candidate beams, or may receive a radio frequency (RF) signal from an external device. The communication module 120 may down-convert an RF signal received from an external device to generate an intermediate frequency signal or a baseband signal. The processor 110 may filter, decode, and / or digitize the intermediate frequency signal or the baseband signal to generate a data signal.

[0063] In one or more examples, the communication module 120 may receive a data signal from the processor 110. The communication module 120 may encode, multiplex, and / or analog-convert the received data signal. The communication module 120 may up-convert an intermediate frequency signal or a baseband signal output from the processor 110 to transmit the intermediate frequency signal or the baseband signal as an RF signal to an external device through a plurality of candidate beams.

[0064] In one or more examples, the sensor 130 may detect an operating state (such as position, power, and temperature) or an external environmental state (such as a user state) of the electronic device 100, and may generate an electrical signal and / or a data value corresponding to the detected state. The sensor 130 may include a gesture sensor, a gyroscope sensor, an air pressure sensor, a magnetic sensor, an acceleration sensor, a grip sensor, a proximity sensor, a color sensor, an infrared (IR) sensor, a biometric sensor, a temperature sensor, a humidity sensor, and / or an illumination sensor. For example, when the position or orientation of the electronic device 100 is changed, the sensor 130 may generate sensing data, and may send the sensing data to the processor 110, in which the amount of change in the orientation of the electronic device 100 is sensed from a reference axis (e.g., a horizontal axis x, a vertical axis y, and a height axis z) using the position of the electronic device 100 before being changed as a reference (e.g., an origin).

[0065] In one or more examples, the memory 140 may include an operating system, and may include a device driver, an executable library, and / or an application and / or process register containing program code. The operating system and the application may be implemented as software elements by executing code and / or instructions by a processor. For example, the memory 140 may store a program in the form of instructions or program code that is executed to perform operations of training the neural network 112 according to an embodiment, inferring a channel state corresponding to the channel characteristic data based on the trained neural network, and updating the beam strategy to correspond to the inferred channel state.

[0066] Figure 4 is a flowchart illustrating a method of operating an electronic device according to one or more embodiments.

[0067] Reference Figure 4 , the adaptive beam management operation method according to the channel state of the electronic device 100 may include operations S110 to S130. Figure 4 The electronic device 100, the channel state classifier 111 and the neural network 112 described are respectively similar to those in reference Figures 1 to 3 The electronic device 100, channel state classifier 111 and neural network 112 are described to correspond.

[0068] In operation S110, the electronic device 100 may generate channel characteristic data based on information measured from a plurality of candidate beams. The channel characteristic data may include information on at least one of RSRP, PDP, and CIR measured from a plurality of candidate beams.

[0069] In operation S120, the electronic device 100 may infer the channel state by applying the channel characteristic data to the neural network 112 of the channel state classifier 111. The electronic device 100 may identify whether the channel information about the channel characteristic data is obtained, and may select a method (hereinafter, referred to as an inference method) for the channel state classifier 111 to infer the channel state based on the identification result. In one or more examples, as described below with reference to Figure 7The channel information may indicate a channel state class having a label value corresponding to the channel characteristic data. The electronic device 100 may infer the channel state corresponding to the channel characteristic data according to the selected inference method by using the neural network 112 of the channel state classifier 111. The electronic device 100 classifies not only the LoS state or the NLoS state (e.g., binary classification) but also the channel states of various communication scenarios (e.g., multi-class classification) through the channel state classifier 111. For example, the channel states in various communication scenarios may include: an NLoS state including a LoS path, an NLoS state not including a LoS path, an NLoS state in which the angle of arrival (AoA) range of the signal received by the electronic device 100 is less than a first threshold, an NLoS state in which the AoA range of the signal received by the electronic device 100 is greater than or equal to the first threshold, an NLoS state in which the number of clusters is less than a second threshold, or an NLoS state in which the number of clusters is greater than or equal to the second threshold. The channel state according to one or more embodiments is not limited to the above examples, and may also include a channel state according to a channel scenario in other wireless communication environments.

[0070] In operation S130, the electronic device 100 may re-form a plurality of candidate beams based on the inferred channel state. The electronic device 100 may re-form a plurality of candidate beams based on the best beam strategy corresponding to the inferred channel state. Figures 8 to 12 A description of the optimal beam strategy corresponding to each channel state. For example, when the channel state is inferred by the channel state classifier 111 as an NLoS state including a LoS path, the electronic device 100 can re-form multiple candidate beams based on narrow beams according to the optimal beam strategy corresponding to the NLoS state including the LoS path.

[0071] Figure 5A An example of a channel state inference operation of an electronic device according to one or more embodiments is shown.

[0072] Reference Figure 5A , the electronic device 100 (refer to Figure 2 ) can input RSRP pattern data into the channel state classifier 111 (refer to Figure 2 ) (or channel state classifier 111 (refer to Figure 2 ) neural network 112 (refer to Figure 2 )) to infer a channel state corresponding to the channel characteristic data. In one or more examples, the RSRP pattern data may represent channel characteristic data including RSRP values ​​measured from a plurality of candidate beams of the electronic device 100 for a signal received from an external device.

[0073] In one or more examples, the processor 110 of the electronic device 100 may control the communication module 120 (see Figure 2 ), beam training is performed based on all or some of the candidate beams in the predefined beam codebook to generate RSRP pattern data. The processor 110 may input the generated RSRP pattern data to the channel state classifier 111.

[0074] In response to the input of RSRP pattern data, the channel state classifier 111 may output a channel state corresponding to the RSRP pattern data (or likelihood according to a channel state class) through a pre-trained neural network 112 based on a deep learning algorithm. For example, when the first RSRP pattern data is input to the channel state classifier 111, the neural network 112 may infer likelihood through a channel state class (e.g., output #1 (501)). In one or more examples, the likelihood according to the channel state class may include soft decision information indicating a probability that the channel state corresponding to the channel characteristic data includes an NLoS state (or LoS state). For example, the channel state classifier 111 may output the LoS likelihood of the channel state corresponding to the channel characteristic data (e.g., RSRP pattern data) to infer the channel state based on the LoS likelihood. For example, the channel state classifier 111 may output the NLoS likelihood of the channel state corresponding to the channel characteristic data (e.g., RSRP pattern data) to infer the channel state based on the NLoS likelihood. In one or more embodiments, the neural network 112 may infer LoS likelihood instead of NLoS likelihood.

[0075] In one or more examples, the channel state classifier 111 may determine whether to obtain channel information (eg, a label value corresponding to the channel characteristic data) from the channel characteristic data, and may select a method for inferring the channel state based on the determination result. Figure 7 A detailed description of the operation of the channel state classifier 111 selecting a channel state inference method is described.

[0076] The channel state classifier 111 may perform multi-class classification of channel states for various communication scenarios (e.g., output #2 (502), output #3 (503), and output #4 (504)), as well as binary classification for LoS state / NLoS state. In one or more examples, the output form of the channel state classifier 111 may include a hard decision form of determining and outputting one of the multiple classes, or a soft decision form of outputting the probability of each of the multiple classes. For example, when the second RSRP pattern data is input to the channel state classifier 111, the neural network 112 may infer one of the channel state classes belonging to output #2 (502) as the channel state corresponding to the second RSRP pattern data. In one or more examples, the channel state class belonging to output #2 (502) may include a LoS state, an NLoS state including a LoS path, and an NLoS state not including a LoS path. For example, when the third RSRP pattern data is input to the channel state classifier 111, the neural network 112 may infer one of the channel state classes belonging to output #3 (503) as a channel state corresponding to the third RSRP pattern data. In one or more examples, the channel state class belonging to output #3 (503) may include a LoS state, an NLoS state with a small angular spread, and an NLoS state with a large angular spread. The NLoS state with a small angular spread may represent an NLoS state in which the AoA range of the signal received by the electronic device 100 is less than the first threshold, and the NLoS state with a large angular spread may represent an NLoS state in which the AoA range of the signal received by the electronic device 100 is greater than or equal to the first threshold. For example, when the fourth RSRP pattern data is input to the channel state classifier 111, the neural network 112 may infer one of the channel state classes belonging to output #4 (504) as a channel state corresponding to the fourth RSRP pattern data. In one or more examples, the channel state class belonging to output #4 (504) may include a LoS state, an NLoS state determined to have a small number of clusters, and an NLoS state determined to have a large number of clusters. The NLoS state determined to have a small number of clusters may represent an NLoS state in which the number of inferred clusters is less than a second threshold, and the NLoS state determined to have a large number of clusters may represent an NLoS state in which the number of inferred clusters is greater than or equal to the second threshold. The cluster may represent an object that affects the direction of a signal received by the electronic device 100.

[0077] In FIG. 5 , the channel state classifier 111 is shown as inferring one of the channel state classes belonging to output #1 (501) to output #4 (504) as the channel state. However, the embodiments of the present disclosure are not limited thereto. The channel state classifier 111 may select at least one of output #1 (501) to output #4 (504) according to the setting, and may infer one of the channel state classes belonging to the selected at least one output as the channel state. For example, when output #2 (502) is selected according to the setting, the channel state classifier 111 may infer one of the classes belonging to output #2 (502) (e.g., LoS state, NLoS state including LoS path, and NLoS state not including LoS path) as the channel state corresponding to the channel characteristic data. For example, when output #2 (502) and output #3 (503) are selected according to the setting, the channel state classifier 111 may infer one of the classes belonging to output #2 (502) and output #3 (503) (e.g., LoS state, NLoS state including a LoS path, NLoS state not including a LoS path, NLoS state with a small angular spread, and NLoS state with a large angular spread) as the channel state corresponding to the channel characteristic data.

[0078] Figure 5B Examples of channel characteristic data according to one or more embodiments are shown.

[0079] Figure 5B RSRP data (or RSRP pattern data) among the channel characteristic data is shown.

[0080] exist Figure 5B In the drawing, it is assumed that the horizontal axis represents the beam index (or angle area index) of the candidate beams of the electronic device 100, and the vertical axis represents RSRP (or signal power) measured from a plurality of candidate beams.

[0081] Reference Figure 5B , RSRP data (or RSRP pattern data) may indicate the presence or absence of a channel path in the region of interest in the spatial domain, and when a channel path exists in the region of interest, RSRP data (or RSRP pattern data) may indicate the power of a signal received through the channel path. For example, RSRP data (or RSRP pattern data) may indicate channel characteristics in the spatial domain. Therefore, the channel state classifier 111 may receive RSRP data (or RSRP pattern data), and may infer a channel state corresponding to the RSRP data (or RSRP pattern data) based on the neural network 112.

[0082] Figure 5C An example of output data of the channel state classifier 111 according to one or more embodiments is shown.

[0083] Figure 5C The likelihood according to the channel state class output by the channel state classifier 111 is shown. Figure 5A In response to the input of the first RSRP pattern data, the channel state classifier 111 may output NLoS likelihood (or LoS likelihood) based on the first RSRP pattern data, and may infer the channel state corresponding to the input first RSRP pattern data based on the output NLoS likelihood (or LoS likelihood). Figure 5C In FIG. 1 , for convenience, it is assumed that the channel state classifier 111 infers the channel state based on the NLoS likelihood.

[0084] Reference Figure 5C , in the NLoS likelihood, it may be noted that a range corresponding to “a probability that the channel state corresponding to the first RSRP pattern data includes a LoS state” is “s”, a range corresponding to “a probability that the channel state corresponding to the first RSRP pattern data includes a first NLoS state (e.g., NLoS state #1)” is “r”, a range corresponding to “a probability that the channel state corresponding to the first RSRP pattern data includes a second NLoS state (e.g., NLoS state #2)” is “q”, and a range corresponding to “a probability that the channel state corresponding to the first RSRP pattern data includes a third NLoS state (e.g., NLoS state #3)” is “p” (p+q+r+s=1). The channel state classifier 111 may infer the first NLoS state (e.g., NLoS state #1) having the widest range “r” among p, q, r, and s as the channel state corresponding to the first RSRP pattern data.

[0085] exist Figure 5C , the channel state classifier 111 is described as inferring the channel state based on NLoS likelihood. However, the embodiments of the present disclosure are not limited thereto, and the channel state classifier 111 according to one or more embodiments may infer the channel state based on LoS likelihood.

[0086] Although for convenience Figure 5C , the likelihood by class (e.g., channel state class) for four channel states is shown, but the embodiments of the present disclosure are not limited thereto, and the channel state classifier 111 according to one or more embodiments may infer the channel state based on the likelihood by channel state class for a larger number of channel states.

[0087] Fig. 6A An example of a channel state inference operation of the electronic device 100 according to one or more embodiments is shown.

[0088] Reference Fig. 6A, the processor 110 of the electronic device 100 (refer to Figure 2 ) can control the communication module 120 (refer to Figure 2 ) performs beam training based on all candidate beams or a partial set of candidate beams in a predefined beam codebook, and measures RSRP including channel characteristic information in the spatial domain and PDP (or CIR) including channel characteristic information in the time domain (at this time, RSRP and PDP (or CIR) may be measured from the same candidate beam, and PDP (or CIR) may not be measured from some candidate beams). The processor 110 may generate two-dimensional channel characteristic data based on the measured information. For example, the processor 110 may generate two-dimensional channel characteristic data by combining RSRP pattern data and PDP data or RSRP pattern data and CIR data measured from a plurality of candidate beams. The processor 110 may input the two-dimensional channel characteristic data into a channel state classifier 111 including a neural network 112.

[0089] In one or more examples, in response to the input of the two-dimensional channel characteristic data, the channel state classifier 111 may output a channel state corresponding to the two-dimensional channel characteristic data by using a pre-trained neural network 112 based on a deep learning algorithm. For example, when the two-dimensional channel characteristic data is input to the channel state classifier 111, the neural network 112 may infer the channel state corresponding to the two-dimensional channel characteristic data.

[0090] In one or more examples, the channel state classifier 111 may determine whether to obtain channel information (eg, a label value corresponding to the channel characteristic data) from the channel characteristic data, and may select a method for inferring the channel state based on the determination result. Figure 7 A detailed description of the operation of the channel state classifier 111 selecting an inference method is described.

[0091] In one or more examples, the channel state classifier 111 may perform multi-class classification of channel states for various communication scenarios, and binary classification for LoS state / NLoS state. In one or more examples, the output form of the channel state classifier 111 may include a hard decision form of determining and outputting one of the multiple classes, or a soft decision form of outputting the probability of each of the multiple classes. In one or more examples, the multiple classes for the channel state may include channel state classes for the following: LoS state, NLoS state including LoS path, NLoS state not including LoS path, NLoS state with small angular spread, NLoS state with large angular spread, NLoS state with a small number of clusters, and NLoS state with a large number of clusters. The NLoS state with small angular spread may represent an NLoS state in which the AoA range of the signal received by the electronic device 100 is less than the first threshold, and the NLoS state with large angular spread may represent an NLoS state in which the AoA range of the signal received by the electronic device 100 is greater than or equal to the first threshold. The NLoS state determined to have a small number of clusters may represent an NLoS state in which the number of inferred clusters is less than a second threshold, and the NLoS state determined to have a large number of clusters may represent an NLoS state in which the number of inferred clusters is greater than or equal to the second threshold. The cluster may represent an object that affects the direction of a signal received by the electronic device 100.

[0092] The channel state classifier 111 according to one or more embodiments may infer channel states (or multiple classes for channel states) in various other communication scenarios / communication environments as well as the above-mentioned channel states (or multiple classes for channel states).

[0093] exist Figure 6B , the channel characteristic data is shown as two-dimensional data. However, the embodiments of the present disclosure are not limited thereto, and the channel characteristic data according to one or more embodiments (e.g., RSRP mode data and PDP data (or RSRP mode data and CIR data)) may be arranged in a one-dimensional form and may be input to the channel state classifier 111 (or the neural network 112). The channel characteristic data according to one or more embodiments (e.g., RSRP mode data and PDP data (or RSRP mode data and CIR data)) may be input to the channel state classifier 111 (or the neural network 112) in a multi-dimensional form.

[0094] Figure 6B Examples of channel characteristic data according to one or more embodiments are shown.

[0095] Figure 6B Shows PDP data (or CIR data) among the channel characteristic data.

[0096] exist Figure 6B In the drawing, assume that the horizontal axis represents a time index (or a resolvable path index), and the vertical axis represents PDP data (or CIR data) measured from a plurality of candidate beams of the electronic device 100.

[0097] Reference Figure 6B , the PDP data (or CIR data) may represent the number of channel paths in the region of interest in the time domain, and when a channel path exists in the region of interest, the PDP data (or CIR data) may represent the power of a signal received through the channel path or the gain of the signal (e.g., the complex gain of the signal). That is, the PDP data (or CIR data) may represent channel characteristics in the time domain. Therefore, the channel state classifier 111 may receive the PDP data (or CIR data), and may infer the channel state corresponding to the PDP data (or CIR data) through the neural network 112.

[0098] Figure 7 is a flowchart illustrating a method of operating the electronic device 100 according to one or more embodiments.

[0099] Figure 7 The electronic device 100 is shown in Figure 4 FIG. 1 is a diagram illustrating an operation of selecting a method of inferring a channel state after generating channel characteristic data in operation S110 of FIG.

[0100] Reference Figure 7 , the operation of the electronic device 100 selecting a method for inferring a channel state may include operations S121 to S126. Figure 7 The electronic device 100, the channel state classifier 111 and the neural network 112 described above can be respectively compared with the reference Figures 1 to 6B The electronic device 100, channel state classifier 111 and neural network 112 are described to correspond.

[0101] In operation S121, the electronic device 100 may determine whether channel information about all channel characteristic data is obtained. In one or more examples, the channel information may represent a label value corresponding to the channel characteristic data (e.g., a channel state class corresponding to the channel characteristic data). For example, an inference method (or inference target) for the channel state classifier 111 (or the neural network 112) to infer the channel state class may be selected based on whether the channel information (or label value) corresponding to the channel characteristic data is available. When the channel information about all channel characteristic data is obtained, the electronic device 100 performs operation S122. If the electronic device 100 determines that the channel information about all channel characteristic data is not obtained, the electronic device 100 may perform operation S123.

[0102] In operation S122, the electronic device 100 may infer the channel state by training the channel state classifier 111 based on the channel information about the channel characteristic data. For example, the channel state classifier 111 of the electronic device 100 may perform binary classification on a new channel based on the training content, and classify the new channel into a LoS state / NLoS state. In one or more examples, the channel state classifier 111 may perform multi-class classification to classify channel states in various scenarios (e.g., an NLoS state including a LoS path and an NLoS state not including a LoS path). The channel state classifier 111 may infer the channel state based on the classification result obtained by the neural network 112.

[0103] In operation S123, the electronic device 100 may perform manifold training through the channel state class. In one or more examples, manifold training may refer to the work of measuring the distance between data items (e.g., channel characteristic data) re-represented in an arbitrary space (e.g., a manifold) to determine the characteristics between the data items (e.g., the similarity between the data items). For example, the neural network 112 may re-represent the channel characteristic data as spatial coordinates in an arbitrary space (e.g., a manifold), and may determine the channel characteristics between the channel characteristic data based on the distance between the spatial coordinates. For example, the electronic device 100 may learn a manifold for a LoS state / NLoS state, or may learn a manifold for channel states in various scenarios.

[0104] In operation S124, the electronic device 100 may determine whether channel information about the partial channel characteristic data is obtained. In one or more examples, the channel information may represent a label value corresponding to the channel characteristic data (e.g., a channel state class corresponding to the channel characteristic data). When the channel information about the partial channel characteristic data is obtained, the electronic device 100 performs operation S125. If the electronic device 100 determines that the channel information about the partial channel characteristic data is not obtained, the electronic device 100 may perform operation S126.

[0105] In operation S125, the electronic device 100 may infer the channel state by inferring the channel information about the remaining channel characteristic data based on the channel information about the partial channel characteristic data. The channel state classifier 111 may infer the channel information (e.g., label value) corresponding to the remaining channel characteristic data based on the channel characteristic data in which the channel information among the channel characteristic data re-represented in an arbitrary space by the neural network 112 exists. For example, the channel state classifier 111 may infer the channel information about the remaining channel characteristic data based on the channel information of the LoS state / NLoS state (binary classification) about the partial channel characteristic data. For example, the channel state classifier 111 may infer the channel information about the remaining channel characteristic data based on the channel information of the channel state in various scenarios (multi-class classification) about the partial channel characteristic data. The channel state classifier 111 may infer the channel state corresponding to the channel characteristic data based on the channel information (including the inferred channel information).

[0106] In operation S126, the electronic device 100 may infer a channel state by predicting channel information of the channel characteristic data based on the distribution of the channel characteristic data. The channel state classifier 111 may predict the channel information of the channel characteristic data based on the distribution of the channel characteristic data (e.g., RSRP and PDP (or CIR)) re-represented in an arbitrary space by the neural network 112. The channel state classifier 111 may infer a channel state corresponding to the channel characteristic data based on the predicted channel information. For example, the channel state classifier 111 may infer a channel state by predicting channel information about a LoS state / NLoS state (binary classification) based on the distribution of the channel characteristic data re-represented in an arbitrary space. For example, the channel state classifier 111 may infer a channel state by predicting channel information about channel states in various scenarios (multi-class classification) based on the distribution of the channel characteristic data re-represented in an arbitrary space.

[0107] Figure 8 is a flowchart illustrating a method of operating the electronic device 100 according to one or more embodiments.

[0108] Figure 8 It is shown Figure 4 Example 800 of operation S130.

[0109] Reference Figure 8 The operation S130 of the electronic device 100 to re-form a plurality of candidate beams based on the channel state may include operations S131 to S135. Figure 8 The electronic device 100, the channel state classifier 111 and the neural network 112 described are respectively similar to those in reference Figures 1 to 7 The electronic device 100, channel state classifier 111 and neural network 112 are described to correspond.

[0110] In operation S131, the electronic device 100 may determine whether the inferred channel state includes the LoS state. When the inferred channel state includes the LoS state, the electronic device 100 may perform operation S132, and when the inferred channel state does not include the LoS state (eg, when the inferred channel state includes the NLoS state), the electronic device 100 may perform operation S133.

[0111] In operation S132, the electronic device 100 may reform the plurality of candidate beams based on the narrow beam. For example, the electronic device 100 may configure the candidate beam set (including the plurality of candidate beams) to include the narrow beam according to the beam strategy corresponding to the LoS state. For example, the electronic device 100 may reform at least one of the plurality of candidate beams into a narrow beam.

[0112] In operation S133, the electronic device 100 may determine whether the inferred channel state includes a LoS path (NLoS state). When the inferred channel state includes an NLoS state including a LoS path, the electronic device 100 may perform operation S134, and if the inferred channel state does not include an NLoS state including a LoS path (i.e., when the inferred channel state includes an NLoS state not including a LoS path), the electronic device 100 may perform operation S135.

[0113] In operation S134, the electronic device 100 may reform the plurality of candidate beams based on the narrow beam. For example, the electronic device 100 may configure the candidate beam set (including the plurality of candidate beams) to include the narrow beam according to the beam strategy corresponding to the NLoS state including the LoS path. For example, the electronic device 100 may reform at least one of the plurality of candidate beams into a narrow beam.

[0114] In operation S135, the electronic device 100 may reform the plurality of candidate beams based on the wide beam. For example, the electronic device 100 may configure the candidate beam set (including the plurality of candidate beams) to include the wide beam according to a beam strategy corresponding to an NLoS state that does not include a LoS path. For example, the electronic device 100 may reform at least one of the plurality of candidate beams into a wide beam.

[0115] Fig. 9 is a flowchart illustrating a method of operating the electronic device 100 according to one or more embodiments.

[0116] Fig. 9 It is shown Figure 4 900 of example of operation S130.

[0117] Reference Fig. 9The operation S130 of the electronic device 100 to reform the beams of the channel state based on the plurality of candidate beams of the channel state may include operations S141 to S145. Fig. 9 The electronic device 100, the channel state classifier 111 and the neural network 112 described are respectively similar to those in reference Figures 1 to 8 The electronic device 100, channel state classifier 111 and neural network 112 are described to correspond.

[0118] In operation S141, the electronic device 100 may determine whether the inferred channel state includes the LoS state. When the inferred channel state includes the LoS state, the electronic device 100 may perform operation S142, and when the inferred channel state does not include the LoS state (ie, when the inferred channel state includes the NLoS state), the electronic device 100 may perform operation S143.

[0119] In operation S142, the electronic device 100 may reform a plurality of candidate beams by activating a sensor-assisted beam management function. For example, the electronic device 100 may reform a plurality of candidate beams according to a beam strategy corresponding to the LoS state by activating a sensor-assisted beam management function. In one or more examples, the sensor-assisted beam management function may represent a function of correcting an existing target beam by generating compensation information according to a change in the position / orientation of the electronic device 100. The existing target beam may represent a beam selected from a plurality of candidate beams based on a previous beam training result. For example, when the position or orientation of the electronic device 100 is changed, the electronic device 100 (or the processor 110 (refer to Figure 2 )) may be based on the slave sensor 130 (refer to Figure 2 ) received sensing data (e.g., data about the amount of change in the orientation of the electronic device 100) to generate compensation information about the existing target beam, and the existing target beam may be corrected based on the generated compensation information. The electronic device 100 may select the corrected target beam as the final target beam, and may perform wireless communication with the external device.

[0120] In operation S143, the electronic device 100 may determine whether the inferred channel state includes a LoS path (NLoS state). When the inferred channel state includes an NLoS state including a LoS path, the electronic device 100 may perform operation S144, and if the inferred channel state does not include an NLoS state including a LoS path (i.e., when the inferred channel state includes an NLoS state not including a LoS path), the electronic device 100 may perform operation S145.

[0121] In operation S144, the electronic device 100 may reform the plurality of candidate beams by activating a sensor-assisted beam management function. For example, the electronic device 100 may reform the plurality of candidate beams by activating a sensor-assisted beam management function according to a beam strategy corresponding to an NLoS state including a LoS path. In one or more examples, the sensor-assisted beam management function may refer to a function of correcting an existing target beam by generating compensation information according to a change in the position / orientation of the electronic device 100. The existing target beam may refer to a beam selected from a plurality of candidate beams based on a previous beam training result.

[0122] In operation S145, the electronic device 100 may reform the plurality of candidate beams by disabling the sensor-assisted beam management function. For example, the electronic device 100 may reform the plurality of candidate beams by disabling the sensor-assisted beam management function according to a beam strategy corresponding to an NLoS state not including a LoS path.

[0123] Fig.10 is a flowchart illustrating a method of operating the electronic device 100 according to one or more embodiments.

[0124] Fig.10 It is shown Figure 4 1000 is a diagram of an example of operation S130 of FIG.

[0125] Reference Fig.10 , the operation S130 in which the electronic device 100 reforms a plurality of candidate beams based on the channel state may include operations S151 to S155. Fig.10 The electronic device 100, the channel state classifier 111 and the neural network 112 described are respectively similar to those in reference Figures 1 to 9 The electronic device 100, channel state classifier 111 and neural network 112 are described to correspond.

[0126] In operation S151, the electronic device 100 may determine whether the inferred channel state includes the LoS state. When the inferred channel state includes the LoS state, the electronic device 100 may perform operation S152, and when the inferred channel state does not include the LoS state (ie, when the inferred channel state includes the NLoS state), the electronic device 100 may perform operation S153.

[0127] In operation S152, the electronic device 100 may reform the plurality of candidate beams based on the narrow beam. For example, the electronic device 100 may configure the candidate beam set (including the plurality of candidate beams) to include the narrow beam according to the beam strategy corresponding to the LoS state. For example, the electronic device 100 may reform at least one of the plurality of candidate beams into a narrow beam.

[0128] In operation S153, the electronic device 100 may determine whether the AoA range of the signal received from the external device is less than the first threshold (NLoS state). The NLoS state in which the AoA range of the signal is less than the first threshold corresponds to the NLoS state with a small angular spread, and the NLoS state in which the AoA range of the signal is the first threshold or more corresponds to the NLoS state with a large angular spread. In the NLoS state in which the AoA range of the received signal is less than the first threshold, the electronic device 100 may perform operation S154, and in the NLoS state in which the AoA range of the received signal is the first threshold or more, the electronic device 100 may perform operation S155.

[0129] In operation S154, the electronic device 100 may reform the plurality of candidate beams based on the first wide beam covering the first angular range. For example, the electronic device 100 may configure the candidate beam set (including the plurality of candidate beams) to include the first wide beam covering the first angular range according to a beam strategy corresponding to an inferred state (e.g., an NLoS state in which the AoA range of the signal is less than a first threshold). For example, the electronic device 100 may reform at least one of the plurality of candidate beams into the first wide beam covering the first angular range.

[0130] In operation S155, the electronic device 100 may reform the plurality of candidate beams based on the second wide beam covering the second angular range. For example, the electronic device 100 may configure the candidate beam set (including the plurality of candidate beams) to include the second wide beam covering the second angular range according to the beam strategy corresponding to the inferred state (i.e., the NLoS state in which the AoA range of the signal is the first threshold or greater). For example, the electronic device 100 may reform at least one of the plurality of candidate beams into the second wide beam covering the second angular range. In one or more examples, the second angular range may be greater than the first angular range.

[0131] Fig.11 is a flowchart illustrating a method of operating the electronic device 100 according to one or more embodiments.

[0132] Fig.11 It is shown Figure 4 Example 1100 of operation S130.

[0133] Reference Fig.11 , the operation S130 in which the electronic device 100 reforms a plurality of candidate beams based on the channel state may include operations S161 to S165. Fig.11 The electronic device 100, the channel state classifier 111 and the neural network 112 described are respectively similar to those in reference Figures 1 to 10 The electronic device 100, channel state classifier 111 and neural network 112 are described to correspond.

[0134] In operation S161, the electronic device 100 may determine whether the inferred channel state includes the LoS state. When the inferred channel state includes the LoS state, the electronic device 100 may perform operation S162, and when the inferred channel state does not include the LoS state (ie, when the inferred channel state includes the NLoS state), the electronic device 100 may perform operation S163.

[0135] In operation S162, the electronic device 100 may reform the plurality of candidate beams based on the narrow beam. For example, the electronic device 100 may configure the candidate beam set (including the plurality of candidate beams) to include the narrow beam according to the beam strategy corresponding to the LoS state. For example, the electronic device 100 may reform at least one of the plurality of candidate beams into a narrow beam.

[0136] In operation S163, the electronic device 100 may determine whether the number of clusters in the direction of the signal is less than the second threshold value (NLoS state). An NLoS state in which the number of clusters in the direction of the signal is less than the second threshold value corresponds to an NLoS state determined to have a small number of clusters, and an NLoS state in which the number of clusters in the direction of the signal is the second threshold value or more corresponds to an NLoS state determined to have a large number of clusters. In the NLoS state in which the number of clusters in the direction of the signal is less than the second threshold value, the electronic device 100 may perform operation S164, and in the NLoS state in which the number of clusters in the direction of the signal is the second threshold value or more, the electronic device 100 may perform operation S165.

[0137] In operation S164, the electronic device 100 may reform a plurality of candidate beams based on a third wide beam including Y1 main lobes. For example, the electronic device 100 may configure a candidate beam set (including a plurality of candidate beams) to include a third wide beam including Y1 main lobes according to a beam strategy corresponding to an inferred state (i.e., an NLoS state in which the number of clusters in the direction of the signal is less than a second threshold). For example, the electronic device 100 may reform at least one of the plurality of candidate beams into a third wide beam including Y1 main lobes.

[0138] In operation S165, the electronic device 100 may reform the plurality of candidate beams based on the fourth wide beam including Y2 main lobes. For example, the electronic device 100 may configure the candidate beam set (including the plurality of candidate beams) to include the fourth wide beam including Y2 main lobes according to the beam strategy corresponding to the inferred state (i.e., the NLoS state in which the number of clusters in the direction of the signal is the second threshold or more). For example, the electronic device 100 may reform at least one of the plurality of candidate beams into the fourth wide beam including Y2 main lobes.

[0139] Fig.12 is a flowchart illustrating a method of operating the electronic device 100 according to one or more embodiments.

[0140] Fig.12 It is shown in Figure 4 FIG. 1 is a diagram illustrating an example of an operation of re-forming a plurality of beams as a channel state of the electronic device 100 is changed after operation S130 of the electronic device 100.

[0141] Reference Fig.12 , the operation of re-forming a plurality of candidate beams as the channel state of the electronic device 100 is changed may include operations S171 and S172. Fig.12 The electronic device 100, the channel state classifier 111 and the neural network 112 described are respectively similar to those in reference Figures 1 to 11 The electronic device 100, channel state classifier 111 and neural network 112 are described to correspond.

[0142] In operation S171, the electronic device 100 may determine whether the channel state inferred by the channel state classifier 111 is changed. When the channel state inferred by the channel state classifier 111 is changed, the electronic device 100 may perform operation S172.

[0143] In operation S172, the electronic device 100 may reform the plurality of candidate beams to correspond to the changed channel state. For example, when the channel state inferred by the channel state classifier 111 is changed from "NLoS state including the LoS path" to "NLoS state not including the LoS path", the electronic device 100 may reform the plurality of candidate beams based on a wide beam (i.e., by invalidating the LoS path) according to a beam strategy corresponding to the NLoS state not including the LoS path.

[0144] Fig.13A is a diagram illustrating a beam strategy corresponding to a channel state inferred by the electronic device 100 according to one or more embodiments.

[0145] According to one or more embodiments, Fig.13AThe representation shows the input of the channel state classifier 111 (e.g., channel characteristic data), the output of the channel state classifier 111 (e.g., channel state), and the beam strategy corresponding to the output of the channel state classifier 111. The input / output of the neural network 112 may correspond to the input / output of the channel state classifier 111.

[0146] Reference Fig.13A , the input of the channel state classifier 111 may include channel characteristic data, and the channel characteristic data may include one of RSRP pattern data, {RSRP pattern data and PDP data} and {RSRP pattern data and CIR data} measured from a plurality of candidate beams. In one or more examples, {RSRP pattern data and PDP data} or {RSRP pattern data and CIR data} may be input to the channel state classifier 111 in a one-dimensional form or a two-dimensional form.

[0147] The output of the channel state classifier 111 is “channel state corresponding to channel characteristic data” inferred by the channel state classifier 111. The channel state classifier 111 may infer and output at least one of the channel classes belonging to output #1 to output #4 as the channel state corresponding to the channel characteristic data.

[0148] Processor 110 (see Figure 2 ) may reform a plurality of candidate beams according to a beam strategy corresponding to an output of the channel state classifier 111 (eg, the inferred channel state).

[0149] In one or more embodiments, in the case of output #1, the channel state classifier 111 may output NLoS (or LoS) likelihood based on the received channel characteristic data (see Figure 5C ), and may finally output a channel state inferred based on the output “NLoS (or LoS) likelihood”. The processor 110 may reform a plurality of candidate beams into a wide beam or a narrow beam according to the finally output channel state (e.g., NLoS state or LoS state). For example, when the finally output channel state is the LoS state, the processor 110 may reform a plurality of candidate beams based on a narrow beam, and when the finally output channel state is the NLoS state, the processor 110 may reform a plurality of candidate beams based on a wide beam.

[0150] In one or more embodiments, in the case of output #2, the channel state classifier 111 may infer / output one of the classes belonging to output #2 (e.g., LoS state, NLoS state including LoS path, and NLoS state not including LoS path) as the channel state based on the received channel characteristic data.

[0151] In one or more embodiments, when the output of the channel state classifier 111 includes the LoS state, the processor 110 may reshape the plurality of candidate beams into narrow beams according to the beam strategy corresponding to the channel state (see Fig.13A Case #1 and Figure 8 ). When the output of the channel state classifier 111 includes an NLoS state including a LoS path, the processor 110 may reform a plurality of candidate beams based on a narrow beam according to a beam strategy corresponding to the channel state. When the output of the channel state classifier 111 includes an NLoS state not including a LoS path, the processor 110 may reform a plurality of candidate beams based on a wide beam according to a beam strategy corresponding to the channel state.

[0152] In one or more embodiments, when the output of the channel state classifier 111 includes the LoS state, the processor 110 may activate the sensor-assisted beam management function according to the beam strategy corresponding to the channel state to re-form a plurality of candidate beams (see Fig.13A Case #2 and Fig. 9 ). When the output of the channel state classifier 111 includes an NLoS state including a LoS path, the processor 110 may activate the sensor-assisted beam management function according to the beam strategy corresponding to the channel state to reform the plurality of candidate beams. When the output of the channel state classifier 111 includes an NLoS state not including a LoS path, the processor 110 may deactivate the sensor-assisted beam management function according to the beam strategy corresponding to the channel state to reform the plurality of candidate beams.

[0153] In one example, when the output of the channel state classifier 111 is changed from "NLoS state including the LoS path" to "NLoS state not including the LoS path", the processor 110 may re-form the plurality of candidate beams into wide beams (refer to Fig.13A Case #3 and Fig.12 ).

[0154] In one or more embodiments, in the case of output #3, the channel state classifier 111 may infer / output one of the classes belonging to output #3 (e.g., a LoS state, an NLoS state with a small angular spread, and an NLoS state with a large angular spread) as the channel state based on the received channel characteristic data. The NLoS state with a small angular spread may represent an NLoS state in which the AoA range of the signal received by the electronic device 100 is less than a first threshold, and the NLoS state with a large angular spread may represent an NLoS state in which the AoA range of the signal received by the electronic device 100 is the first threshold or greater. When the output of the channel state classifier 111 includes the LoS state, the processor 110 may reform a plurality of candidate beams based on narrow beams according to a beam strategy corresponding to the channel state (refer to Fig.10 ). When the output of the channel state classifier 111 includes an NLoS state with a small angular spread (i.e., an NLoS state in which the AoA range of the received signal is less than a first threshold), the processor 110 may reform a plurality of candidate beams based on a wide beam covering an X1 angular range according to a beam strategy corresponding to the channel state. When the output of the channel state classifier 111 includes an NLoS state with a large angular spread (i.e., an NLoS state in which the AoA range of the received signal is a first threshold or greater), the processor 110 may reform a plurality of candidate beams based on a wide beam covering an X2 (X2>X1) angular range according to a beam strategy corresponding to the channel state.

[0155] In one or more embodiments, in the case of output #4, the channel state classifier 111 may infer / output one of the classes belonging to output #4 (e.g., a LoS state, an NLoS state with a small number of clusters, and an NLoS state with a large number of clusters) as the channel state based on the received channel characteristic data. The NLoS state determined to have a small number of clusters may represent an NLoS state in which the number of clusters is less than a second threshold, and the NLoS state determined to have a large number of clusters may represent an NLoS state in which the number of clusters is greater than or equal to the second threshold. A cluster may represent an object that affects the direction of a signal received by the electronic device 100. When the output of the channel state classifier 111 includes the LoS state, the processor 110 may reform a plurality of candidate beams based on narrow beams according to a beam strategy corresponding to the channel state (refer to Fig.11). When the output of the channel state classifier 111 includes an NLoS state determined to have a small number of clusters (e.g., an NLoS state in which the number of clusters is less than the second threshold), the processor 110 may reform a plurality of candidate beams based on a wide beam including Y1 main lobes according to a beam strategy corresponding to the channel state. When the output of the channel state classifier 111 includes an NLoS state determined to have a large number of clusters (e.g., an NLoS state in which the number of clusters is the second threshold or greater), the processor 110 may reform a plurality of candidate beams based on a wide beam including Y2 (Y2>Y1) main lobes according to a beam strategy corresponding to the channel state.

[0156] The apparatus and method according to the embodiment may reduce beam training overhead by performing adaptive beam management (eg, reforming a plurality of candidate beams) according to an inferred channel state, and may improve communication performance of the apparatus by applying a beam strategy corresponding to the channel state.

[0157] Fig. 13B An example of candidate beams re-formed by the electronic device 100 according to one or more embodiments is shown.

[0158] Fig. 13B A wide beam 1301 and a narrow beam 1302 are shown as Fig.13A Example of candidate beams reformed using the beam strategy of FIG.

[0159] Reference Fig. 13B In case (a), the processor 110 (refer to Figure 2 ) may reform multiple candidate beams based on the wide beam 1301 according to a beam strategy corresponding to the inferred channel state (e.g., an NLoS state that does not include a LoS path). In case of (b), the processor 110 may reform multiple candidate beams based on the narrow beam 1302 according to a beam strategy corresponding to the inferred channel state (e.g., an NLoS state that includes a LoS path). Fig. 13B As shown in , in one or more examples, the narrow beam 1302 may include a first beam width W1, and the wide beam 1301 may include a second beam width W2, where W2>W1.

[0160] Fig.14A An example of a beam management operation based on a channel state of the electronic device 100 according to one or more embodiments is shown.

[0161] Reference Fig.14AIn case (a), the electronic device 100 may apply the channel characteristic data measured from the plurality of candidate beams for the signal received from the external device to the channel state classifier 111 to infer the channel state of the channel receiving the signal as the "LoS state". The electronic device 100 may control the communication module 120 (refer to Figure 2 ) Re-form multiple candidate beams based on narrow beams according to the beam strategy corresponding to the "LoS state".

[0162] In case of (b), the electronic device 100 may apply the channel characteristic data measured from the plurality of candidate beams for the signal received from the external device to the channel state classifier 111 to infer the channel state of the channel receiving the signal as the “NLoS state including the LoS path”. The electronic device 100 may control the communication module 120 (refer to Figure 2 ) Re-form a plurality of candidate beams based on narrow beams according to a beam strategy corresponding to an “NLoS state including a LoS path”.

[0163] In case of (c), the electronic device 100 may apply the channel characteristic data measured from the plurality of candidate beams for the signal received from the external device to the channel state classifier 111 to infer the channel state of the channel receiving the signal as “NLoS state not including LoS path”. The electronic device 100 may control the communication module 120 (refer to Figure 2 ) Re-form a plurality of candidate beams based on a wide beam according to a beam strategy corresponding to an “NLoS state not including a LoS path”.

[0164] Fig. 14B An example of a beam training operation using candidate beams reformed for each channel state according to one or more embodiments is shown.

[0165] Specifically, Fig. 14B is used to explain Fig.14A FIG. 5 is a diagram of a beam training operation for reforming multiple candidate beams in the case of (a).

[0166] Reference Fig. 14B, the electronic device 100 may reform a plurality of candidate beams based on a narrow beam according to a beam strategy corresponding to an inferred channel state (e.g., a LoS state). For example, the electronic device 100 may reform the first candidate beam NB1 to the seventh candidate beam NB7 into narrow beams. The electronic device 100 may perform beam training based on the reformed plurality of candidate beams. For example, the electronic device 100 may generate RSRP pattern data (e.g., at time point t1) by measuring RSRP from each of the reformed plurality of candidate beams (e.g., the first candidate beam NB1 to the seventh candidate beam NB7). The electronic device 100 may select the seventh candidate beam NB7 having "-95.16" as the highest value in the RSRP pattern data as a target beam. The electronic device 100 may perform wireless communication with an external device by using the seventh candidate beam NB7.

[0167] Fig. 14C An example of a beam training operation using candidate beams reformed for each channel state according to one or more embodiments is shown.

[0168] Specifically, Fig. 14C is used to explain Fig.14A (c) is a diagram illustrating a beam training operation for reforming multiple candidate beams.

[0169] Reference Fig. 14C , the electronic device 100 may reform at least one of the multiple candidate beams into a wide beam based on a wide beam according to a beam strategy corresponding to an inferred channel state (e.g., an NLoS state that does not include a LoS path). For example, the electronic device 100 may reform the first candidate beam NB1 to the fourth candidate beam NB4 among the multiple candidate beams into narrow beams, and reform the fifth candidate beam WB1 to the seventh candidate beam WB3 among the multiple candidate beams into wide beams. The electronic device 100 may perform beam training based on the reformed multiple candidate beams. For example, the electronic device 100 may generate RSRP pattern data by measuring RSRP from each of the reformed multiple candidate beams (e.g., the first candidate beam NB1 to the seventh candidate beam WB3). The electronic device 100 may select the sixth candidate beam WB2 having "-90.49" as the highest value in the RSRP pattern data as a target beam. The electronic device 100 may perform wireless communication with an external device by using the sixth candidate beam WB2.

[0170] Fig.15 is a schematic block diagram of a wireless communication system 1500 according to one or more embodiments.

[0171] Fig.15 The electronic device 1501 can be used with reference to Figure 1 to Figure 14 corresponds to the electronic device 100 described in Fig.15 The processor 1520 may be used with reference to Figure 1 to Figure 1 4 corresponds to the processor 110 described above. Fig.15 The memory 1530 can be used with Figure 3 The memory 140 corresponds to, Fig.15 The sensor module 1576 can be used with Figure 3 The sensor 130 corresponds to, and Fig.15 The communication module 1590 can be used with Figure 3 Corresponding to the communication module 120.

[0172] Reference Fig.15 In the wireless communication system 1500, the electronic device 1501 may communicate with another electronic device 1502 through a first network 1598 (such as a short-range wireless communication network), or may communicate with another electronic device 1504 and / or a server 1508 through a second network 1599 (such as a long-range wireless communication network). The electronic device 1501 may communicate with the electronic device 1504 through the server 1508. The electronic device 1501 may include a processor 1520, a memory 1530, an input device 1550, a sound output device 1555, a display device 1560, an audio module 1570, a sensor module 1576, an interface 1577, a connection terminal 1578, a haptic module 1579, a camera module 1580, a power management module 1588, a battery 1589, a communication module 1590, a user identification module 1596, and / or an antenna module 1597. In the electronic device 1501, some of the components (such as the display device 1560) may be omitted, or other components may be added. Some of the components may be implemented as one integrated circuit. For example, a sensor module 1576 (such as a fingerprint sensor, an iris sensor, or an illumination sensor) may be embedded in the display device 1560 (such as a display).

[0173] The processor 1520 may execute software (such as a program 1540) to control one or more other components (such as hardware components and software components) of the electronic device 1501 connected to the processor 1520, and may perform various data processing or operations. As part of the data processing or operation, the processor 1520 may load instructions and / or data received from other components (such as a sensor module 1576 and a communication module 1590) into the volatile memory 1532, may process the instructions and / or data stored in the volatile memory 1532, and may store the resultant data in the non-volatile memory 1534. The processor 1520 may include a main processor 1521 (such as a central processing unit (CPU) or an application processor (AP)) and an auxiliary processor 1523 (such as a graphics processor (GPU), an image signal processor, a sensor hub processor, or a communication processor), and the auxiliary processor 1523 may operate independently of the main processor 1521 or operate together with the main processor 1521. The auxiliary processor 1523 may use less power than the main processor 1521 and may perform dedicated functions.

[0174] The auxiliary processor 1523 may control functions and / or states related to some of the components of the electronic device 1501 (such as the display device 1560, the sensor module 1576, and the communication module 1590) on behalf of the main processor 1521 when the main processor 1521 is in an inactive state (such as a sleep state), or control functions and / or states related to some of the components of the electronic device 1501 (such as the display device 1560, the sensor module 1576, and the communication module 1590) together with the main processor 1521 when the main processor 1521 is in an active state (such as an application execution state). The auxiliary processor 1523 (such as an image signal processor or a communication processor) may be implemented as a part of other function-related components (such as the camera module 1580 and the communication module 1590).

[0175] In one or more embodiments, the processor 1520 may apply channel characteristic data (e.g., RSRP pattern data and PDP data (or CIR data)) measured from multiple candidate beams of the electronic device 1501 for a signal received from an external device to the channel state classifier 111 (or the neural network 112) to infer a channel state corresponding to the channel characteristic data.

[0176] In one or more embodiments, the processor 1520 may perform various beam managements based on the inferred channel state. For example, the processor 1520 may re-form multiple candidate beams according to a beam strategy corresponding to the inferred channel state. The processor 1520 may select a target beam by performing beam training on the re-formed multiple candidate beams. The processor 1520 may perform wireless communication with an external device by using the selected target beam.

[0177] The processor 1520 may re-form a plurality of candidate beams by activating a sensor-assisted beam management function according to a beam strategy corresponding to a channel state. For example, when the position / orientation of the electronic device 1501 is changed after an existing target beam is selected, the processor 1520 may activate / deactivate the sensor-assisted beam management function according to the inferred channel state to correct the existing target beam based on the sensing data of the sensor module 1576 (e.g., the amount of change in the orientation of the electronic device 1501), and may select the corrected beam as the final target beam for performing wireless communication with an external device.

[0178] The memory 1530 may store various data required by components of the electronic device 1501, such as the processor 1520 and the sensor module 1576. The data may include, for example, input data and / or output data for software, such as the program 1540, and instructions related thereto. The memory 1530 may include a volatile memory 1532 and / or a nonvolatile memory 1534. The nonvolatile memory 1532 may include an internal memory 1536 fixedly installed in the electronic device 1501 and a removable external memory 1538.

[0179] The program 1540 may be stored as software in the memory 1530 , and may include an operating system 1542 , middleware 1544 , and / or an application 1546 .

[0180] The input device 1550 may receive instructions and / or data to be used for components (such as the processor 1520) of the electronic device 1501 from outside (such as a user) of the electronic device 1501. The input device 1550 may include a microphone, a mouse, a keyboard, and / or a digital pen (such as a stylus).

[0181] The sound output device 1555 can output a sound signal to the outside of the electronic device 1501. The sound output device 1555 may include a speaker and / or a receiver. The speaker may be used for general purposes (such as multimedia playback or recording playback), and the receiver may be used to receive incoming calls. The receiver may be integrated as part of the speaker, or implemented as a separate independent device.

[0182] The display device 1560 may visually provide information to the outside of the electronic device 1501. The display device 1560 may include a display, a hologram device or a projector, and a control circuit for controlling these devices. The display device 1560 may include a touch circuit system configured to detect a touch and / or a sensor circuit (such as a pressure sensor) configured to measure the strength of the force generated by the touch.

[0183] The audio module 1570 can convert sound into an electrical signal, or vice versa, can convert an electrical signal into sound. The audio module 1570 can obtain sound through the input device 1550, or can output sound through a speaker and / or earphone directly or wirelessly connected to the sound output device 1555 and / or another electronic device (such as, electronic device 1502) of the electronic device 1501.

[0184] The sensor module 1576 may detect an operating state (such as power and temperature) or an external environment state (such as a user state) of the electronic device 1501, and may generate an electrical signal and / or a data value corresponding to the detected state. The sensor module 1576 may include a gesture sensor, a gyroscope sensor, an air pressure sensor, a magnetic sensor, an acceleration sensor, a grip sensor, a proximity sensor, a color sensor, an infrared (IR) sensor, a biometric sensor, a temperature sensor, a humidity sensor, and / or an illumination sensor. For example, the sensor module 1576 may sense a change in the position or orientation of the electronic device 1501.

[0185] The interface 1577 may support one or more specified protocols that may be used to directly or wirelessly connect the electronic device 1501 to another electronic device, such as the electronic device 1502. The interface 1577 may include a high-definition multimedia interface (HDMI), a universal serial bus (USB) interface, a secure digital (SD) card interface, and / or an audio interface.

[0186] The connection terminal 1578 may include a connector through which the electronic device 1501 may be physically connected to another electronic device (such as the electronic device 1502). The connection terminal 1578 may include an HDMI connector, a USB connector, an SD card connector, and / or an audio connector (such as a headphone connector).

[0187] The haptic module 1579 may convert the electric signal into mechanical stimulation (such as vibration and movement) or electric stimulation that the user can sense through tactile or kinesthetic sense. The haptic module 1579 may include a motor, a piezoelectric element, and / or an electric stimulation device.

[0188] The camera module 1580 can capture still images and moving images. The camera module 1580 may include an image acquisition device, an additional lens assembly, an image signal processor, and / or a flash. The lens assembly included in the camera module 1580 may collect light emitted from an object as a target of image capture.

[0189] The power management module 1588 may manage power supplied to the electronic device 1501. The power management module 1588 may be implemented as a part of a power management integrated circuit (PMIC).

[0190] The battery 1589 may supply power to the components of the electronic device 1501. The battery 1589 may include a non-rechargeable primary battery, a rechargeable secondary battery, and / or a fuel cell.

[0191] The communication module 1590 may support the establishment of a direct (wired) communication channel and / or a wireless communication channel between the electronic device 1501 and another electronic device (electronic device 1502, electronic device 1504 or server 1508), and communication through the established communication channel. The communication module 1590 may include one or more communication processors that operate independently of the processor 1520 (such as, AP) and support direct communication and / or wireless communication. The communication module 1590 may include a wireless communication module 1592 (such as, a cellular communication module, a near field communication module or a global navigation satellite system (GNSS) communication module) and / or a wired communication module 1594 (such as, a local area network (LAN) communication module or a power line communication module). Among these communication modules, the corresponding communication module can communicate with other electronic devices through a first network 1598 (a short-range communication network (such as, Bluetooth, WiFi Direct or Infrared Data Association (IrDA))) or a second network 1599 (a long-range communication network (such as, a cellular network, the Internet or a computer network (such as, a LAN or a wide area network (WAN)))). These various types of communication modules may be integrated into one component (such as a single chip) or may be implemented as a plurality of separate components (such as a plurality of chips). The wireless communication module 1592 may verify and authenticate the electronic device 1501 in a communication network (such as a first network 1598 and / or a second network 1599) by using user information (such as an International Mobile Subscriber Identity (IMSI)) stored in the user identification module 1596. For example, the communication module 1590 may perform wireless communication with an external device based on a target beam selected by performing beam training on a plurality of re-formed candidate beams by the processor 1520.

[0192] The antenna module 1597 may send signals and / or power to the outside (such as other electronic devices) or receive signals and / or power from the outside (such as other electronic devices). The antenna may include a radiator including a conductive pattern formed on a substrate (such as a printed circuit board (PCB)). The antenna module 1597 may include one or more antennas. When multiple antennas are included, an antenna suitable for a communication method used in a communication network (such as the first network 1598 and / or the second network 1599) may be selected from the multiple antennas by the communication module 1590. Signals and / or power may be sent or received between the communication module 1590 and other electronic devices through the selected antenna. In addition to the antenna, other components (such as a radio frequency integrated circuit (RFIC)) may be included as part of the antenna module 1597.

[0193] Some of the components are connected to each other, and may exchange signals such as commands and data through a communication method between peripheral devices such as a bus, a general purpose input and output (GPIO), a serial peripheral interface (SPI), and a mobile industry processor interface (MIPI).

[0194] Instructions or data may be sent or received between the electronic device 1501 and the external electronic device 1504 through the server 1508 connected to the second network 1599. The other electronic devices 1502 and 1504 may be devices of the same or different types as the electronic device 1501. All or some of the operations performed by the electronic device 1501 may be performed by one or more of the other electronic devices 1502 and 1504 and the server 1508. For example, when the electronic device 1501 needs to perform a specific function or service, the electronic device 1501 may require one or more electronic devices to perform part or all of the function or service instead of performing the function or service. The one or more electronic devices that receive the request may perform additional functions or services related to the request, and the results of the execution may be sent to the electronic device 1501. To this end, cloud computing, distributed computing and / or client-server computing technologies may be used.

[0195] While embodiments of the present disclosure have been particularly shown and described with reference to embodiments thereof, it will be understood that various changes in form and details may be made therein without departing from the spirit and scope of the appended claims.

Claims

1. An electronic device, comprising: Memory; A communication module comprising an antenna array configured to form a plurality of candidate beams; as well as A processor, operably connected to the communication module and the memory, the processor comprising a channel state classifier based on a neural network, The processor is configured as follows: generating channel characteristic data based on information measured from the plurality of candidate beams for a signal received from a base station, inputting the channel characteristic data into a channel state classifier to infer the channel state between the electronic device and the base station, and The plurality of candidate beams are re-formed based on the inferred channel state.

2. The electronic device according to claim 1, wherein: The channel characteristic data includes information on at least one of reference signal received power, power delay profile, and channel impulse response.

3. The electronic device according to claim 1, wherein: The processor is also configured to: determining whether to obtain channel information from channel characteristic data, and Based on the result of determining whether the channel information is obtained from the channel characteristic data, a method for the channel state classifier to infer the channel state is selected.

4. The electronic device according to claim 3, wherein: The processor is also configured to: Based on determining to obtain channel information corresponding to all channel characteristic data among the channel characteristic data from the channel characteristic data, a method for classifying the channel class based on the channel information is selected; Based on determining that channel information corresponding to all of the channel characteristic data among the channel characteristic data is not obtained from the channel characteristic data, selecting a method of predicting channel information corresponding to the channel characteristic data based on distribution of the channel characteristic data; and Based on determining to obtain channel information corresponding to part of the channel characteristic data from the channel characteristic data, a method of inferring channel information corresponding to the remaining channel characteristic data is selected based on the channel information corresponding to the part of the channel characteristic data.

5. The electronic device according to claim 1, wherein: The processor is further configured to: infer the channel state as one of a line-of-sight state and a non-line-of-sight state, Based on inferring the channel state as a line-of-sight state, reforming the plurality of candidate beams based on a narrow beam having a first beamwidth, and Based on inferring the channel state as a non-line-of-sight state, the plurality of candidate beams are re-formed based on a wide beam having a second beamwidth, the second beamwidth being greater than the first beamwidth.

6. The electronic device according to claim 1, wherein: The processor is further configured to infer the channel state as one of a line-of-sight state, a non-line-of-sight state including a line-of-sight path, and a non-line-of-sight state not including a line-of-sight path, Based on inferring the channel state as a line-of-sight state, reforming the plurality of candidate beams based on a narrow beam having a first beamwidth, Based on inferring the channel state as a non-line-of-sight state including a line-of-sight path, reforming the plurality of candidate beams based on the narrow beams, and Based on inferring the channel state as a non-line-of-sight state that does not include a line-of-sight path, the plurality of candidate beams are reformed based on a wide beam having a second beamwidth, the second beamwidth being greater than the first beamwidth.

7. The electronic device according to claim 6, wherein: The processor is further configured to reform the plurality of candidate beams based on the wide beam based on determining that a channel state is changed from a non-line-of-sight state including a line-of-sight path to a non-line-of-sight state not including a line-of-sight path.

8. The electronic device according to claim 1, wherein: The electronic device further comprises: at least one sensor, The processor is further configured to: inferring a channel state as one of a line-of-sight state, a non-line-of-sight state including a line-of-sight path, and a non-line-of-sight state not including a line-of-sight path, Based on inferring the channel state as a line-of-sight state, activating a sensor-assisted beam management function to re-form the plurality of candidate beams, Based on inferring the channel state as a non-line-of-sight state including a line-of-sight path, activating a sensor-assisted beam management function to reform the plurality of candidate beams, and Based on inferring the channel state as a non-line-of-sight state that does not include a line-of-sight path, deactivating a sensor-assisted beam management function to reform the plurality of candidate beams, and The sensor-assisted beam management function corresponds to a function of correcting the plurality of candidate beams based on position change data sensed by the at least one sensor in response to a change in the position of the electronic device.

9. The electronic device according to claim 1, wherein: The processor is also configured to: inferring the channel state as one of a line-of-sight state, a non-line-of-sight state in which the angle-of-arrival range of the signal is less than a first threshold, and a non-line-of-sight state in which the angle-of-arrival range of the signal is greater than or equal to the first threshold, Based on inferring the channel state as a line-of-sight state, reforming the plurality of candidate beams based on narrow beams, Based on inferring the channel state as a non-line-of-sight state in which the arrival angle range of the signal is less than a first threshold, reforming the plurality of candidate beams based on a first wide beam covering a first angle range, and Based on inferring the channel state as a non-line-of-sight state in which the arrival angle range of the signal is greater than or equal to a first threshold, the plurality of candidate beams are reformed based on a second wide beam covering a second angle range, the second angle range being greater than the first angle range.

10. The electronic device according to claim 1, wherein: The processor is also configured to: inferring the channel state as one of a line-of-sight state, a non-line-of-sight state in which the number of clusters is less than a second threshold, and a non-line-of-sight state in which the number of clusters is greater than or equal to the second threshold, Based on inferring the channel state as a line-of-sight state, reforming the plurality of candidate beams based on narrow beams, Based on inferring the channel state as a non-line-of-sight state in which the number of clusters is less than a second threshold, reforming the plurality of candidate beams based on a third wide beam including a predetermined number of main lobes, and reforming the plurality of candidate beams based on a fourth wide beam including more main lobes than the predetermined number based on inferring the channel state as a non-line-of-sight state in which the number of clusters is greater than or equal to a second threshold, Therein, a cluster represents an object that affects the direction of the signal.

11. A method for operating an electronic device, the method comprising: generating channel characteristic data based on information measured from a plurality of candidate beams for a signal received from a base station; Applying the channel characteristic data to a neural network based on a deep learning algorithm to infer the channel state between the electronic device and the base station; as well as The plurality of candidate beams are re-formed based on the inferred channel state.

12. The method according to claim 11, wherein: The channel characteristic data includes information on at least one of reference signal received power, power delay profile, and channel impulse response.

13. The method according to claim 11, further comprising: determining whether to obtain channel information from channel characteristic data; as well as Based on the result of determining whether the channel information is obtained from the channel characteristic data, a method for the neural network to infer the channel state is selected.

14. The method according to claim 13, wherein: The steps of selecting a method for inferring channel status include: Based on determining to obtain channel information corresponding to all channel characteristic data among the channel characteristic data from the channel characteristic data, a method for classifying the channel class based on the channel information is selected; Based on determining that channel information corresponding to all channel characteristic data among the channel characteristic data is not obtained from the channel characteristic data, selecting a method of predicting channel information corresponding to the channel characteristic data based on distribution of the channel characteristic data; and Based on determining that channel information corresponding to part of the channel characteristic data is obtained from the channel characteristic data, a method of inferring channel information corresponding to remaining channel characteristic data based on the channel information corresponding to the part of the channel characteristic data is selected.

15. The method according to claim 11, wherein: The step of inferring the channel state comprises: inferring the channel state as one of a line-of-sight state and a non-line-of-sight state, and The step of re-forming the multiple candidate beams includes: Based on inferring the channel state as a line-of-sight state, reforming the plurality of candidate beams based on a narrow beam having a first beamwidth; and Based on inferring the channel state as a non-line-of-sight state, the plurality of candidate beams are re-formed based on a wide beam having a second beamwidth, the second beamwidth being greater than the first beamwidth.

16. The method according to claim 11, wherein: The step of inferring the channel state includes inferring the channel state as one of a line-of-sight state, a non-line-of-sight state including a line-of-sight path, and a non-line-of-sight state not including a line-of-sight path, and The step of re-forming the multiple candidate beams includes: reforming the plurality of candidate beams based on narrow beams having a first beamwidth based on inferring the channel state as a line-of-sight state; reforming the plurality of candidate beams based on the narrow beams based on inferring the channel state as a non-line-of-sight state including a line-of-sight path; and Based on inferring the channel state as a non-line-of-sight state that does not include a line-of-sight path, the plurality of candidate beams are reformed based on a wide beam having a second beamwidth, the second beamwidth being greater than the first beamwidth.

17. The method according to claim 11, wherein: The step of inferring the channel state includes inferring the channel state as one of a line-of-sight state, a non-line-of-sight state including a line-of-sight path, and a non-line-of-sight state not including a line-of-sight path, The step of re-forming the multiple candidate beams includes: Based on inferring the channel state as a line-of-sight state, activating a sensor-assisted beam management function to re-form the plurality of candidate beams; Based on inferring the channel state as a non-line-of-sight state including a line-of-sight path, activating a sensor-assisted beam management function to reform the plurality of candidate beams; and Based on inferring the channel state as a non-line-of-sight state that does not include a line-of-sight path, deactivating a sensor-assisted beam management function to reform the plurality of candidate beams, and The sensor-assisted beam management function corresponds to a function of correcting the plurality of candidate beams based on sensing data regarding a change in a position of the electronic device.

18. The method according to claim 11, wherein: The step of inferring the channel state comprises: inferring the channel state as one of a line-of-sight state, a non-line-of-sight state in which the angle-of-arrival range of the signal is less than a first threshold, and a non-line-of-sight state in which the angle-of-arrival range of the signal is greater than or equal to the first threshold, and The step of re-forming the multiple candidate beams includes: reforming the plurality of candidate beams based on narrow beams based on inferring the channel state as a line-of-sight state; Based on inferring the channel state as a non-line-of-sight state in which the arrival angle range of the signal is less than a first threshold, reforming the plurality of candidate beams based on a first wide beam covering a first angle range; and Based on inferring the channel state as a non-line-of-sight state in which the arrival angle range of the signal is greater than or equal to a first threshold, the plurality of candidate beams are reformed based on a second wide beam covering a second angle range, the second angle range being greater than the first angle range.

19. The method according to claim 11, wherein: The step of inferring the channel state comprises: inferring the channel state as one of a line-of-sight state, a non-line-of-sight state in which the number of clusters is less than a second threshold, and a non-line-of-sight state in which the number of clusters is greater than or equal to the second threshold, The step of re-forming the multiple candidate beams includes: reforming the plurality of candidate beams based on narrow beams based on inferring the channel state as a line-of-sight state; reforming the plurality of candidate beams based on a third wide beam including a predetermined number of main lobes based on inferring the channel state as a non-line-of-sight state in which the number of clusters is less than a second threshold; and reforming the plurality of candidate beams based on a fourth wide beam including more main lobes than the predetermined number based on inferring the channel state as a non-line-of-sight state in which the number of clusters is greater than or equal to a second threshold, Therein, a cluster represents an object that affects the direction of the signal.

20. A method of operating an electronic device, the method comprising: forming a plurality of candidate beams for performing wireless communication with an external device; generating first channel characteristic data representing channel characteristics in the spatial domain and second channel characteristic data representing channel characteristics in the time domain based on information measured from the plurality of candidate beams; applying data obtained by combining the first channel characteristic data and the second channel characteristic data in a two-dimensional form to a pre-trained channel state inference model to infer a channel state between the electronic device and the external device; as well as reforming the plurality of candidate beams based on the inferred channel state, wherein the first channel characteristic data includes information about reference signal received powers measured from the plurality of candidate beams, The second channel characteristic data includes information about one of a power delay profile and a channel impulse response measured from the plurality of candidate beams.

Images