Electronic device for supporting online training of neural network for wireless communication and operating method thereof

Abstract

According to an embodiment, an electronic device may be provided, comprising: a plurality of antennas; at least one processor; and a memory for storing at least one instruction. The at least one processor may be configured to receive signals from an external electronic device via at least one of the plurality of antennas. The at least one processor may be configured to: identify whether information associated with characteristics of a wireless communication channel identified from the received signals satisfies a condition for identifying the reliability of training data. The at least one processor may be configured to: perform online training of an artificial neural network for estimating the characteristics of a wireless communication channel corresponding to a cell, based on the information identified as satisfying the condition, using the information associated with the characteristics of the wireless communication channel. The at least one processor may be configured to: estimate the characteristics of the wireless communication channel based on information output from the online-trained artificial neural network by inputting signals received via at least some of the plurality of antennas into the online-trained artificial neural network.

Description

Technical Field

[0001] This disclosure relates to an electronic device for supporting online training of neural networks for wireless communication and a method of operating the same. Background Technology

[0002] Communication systems are constantly evolving to support higher data rates, such as those of 5G communication systems, to meet the demands of wireless data services. 5G communication systems are being considered for implementation in millimeter-wave (mmWave) bands (e.g., the 60GHz band) to achieve data rates approximately 10 times higher than existing 4G communication systems.

[0003] Recently, interest in artificial intelligence (AI) has been increasing, and significant advancements in computing power and the emergence of deep learning technologies have substantially improved the accuracy of AI. Therefore, research is actively underway to utilize AI technologies across various fields. In wireless mobile communication systems, various studies on artificial neural networks are also being conducted. Artificial neural networks can be trained offline. Offline training is a method of training neural networks using pre-acquired data or learning data from a link-level simulator. In the receiver performance of mobile communication systems, channel estimation between the base station and the terminal is one of the key factors. To improve the performance of channel estimation, online training based on real-world field data reflecting the radio environment of the cell is required.

[0004] The information above is provided as relevant technical information to aid in understanding this document. The foregoing is neither claimed to be prior art related to this document nor permitted to be used to determine prior art. Summary of the Invention

[0005] Technical solution

[0006] According to an embodiment, an electronic device may include a plurality of antennas. The electronic device may include at least one processor. The electronic device may include a memory configured to store at least one instruction. The at least one processor may be configured to receive signals from an external electronic device via at least one of the plurality of antennas. The at least one processor may be configured to: identify whether information associated with characteristics of a wireless communication channel identified from the received signals satisfies a condition for identifying the reliability of training data. The at least one processor may be configured to: perform online training of an artificial neural network for estimating the characteristics of a wireless communication channel corresponding to a cell, using the information associated with the characteristics of the wireless communication channel, based on the condition that the identified information associated with the characteristics of the wireless communication channel satisfies the condition. The at least one processor may be configured to: estimate the characteristics of the wireless communication channel based on information output from the online-trained artificial neural network by inputting signals received via at least a portion of the plurality of antennas into the online-trained artificial neural network.

[0007] An embodiment provides a method for estimating the characteristics of a wireless communication channel by an electronic device. The method may include: receiving signals from an external electronic device through at least one of a plurality of antennas of the electronic device. The method may include: identifying whether information associated with the characteristics of the wireless communication channel identified from the received signals satisfies conditions for identifying the reliability of training data. The method may include: performing online training of an artificial neural network for estimating the characteristics of a wireless communication channel corresponding to a cell, using the information associated with the characteristics of the wireless communication channel, based on the satisfaction of the conditions. The method may include: estimating the characteristics of the wireless communication channel based on information output from the online-trained artificial neural network by inputting signals received through at least a portion of the plurality of antennas.

[0008] An embodiment may provide a storage medium storing at least one instruction readable by a computer. When executed by at least one processor of an electronic device, the at least one instruction can cause the electronic device to perform at least one operation. The at least one operation may include: receiving a signal from an external electronic device through at least one of a plurality of antennas of the electronic device. The at least one operation may include: identifying whether information associated with characteristics of a wireless communication channel identified from the received signal satisfies conditions for identifying the reliability of training data. The at least one operation may include: performing online training of an artificial neural network for estimating the characteristics of a wireless communication channel corresponding to a cell, using the information associated with the characteristics of the wireless communication channel, based on the satisfied condition. The at least one operation may include: estimating the characteristics of the wireless communication channel based on information output from the online-trained artificial neural network by inputting signals received through at least a portion of the plurality of antennas. Attached Figure Description

[0009] Figure 1 This is a block diagram of an electronic device in a network environment according to various embodiments.

[0010] Figure 2a It is a block diagram of an electronic device for supporting conventional network communication and 5G network communication according to various embodiments;

[0011] Figure 2b It is a block diagram of an electronic device for supporting conventional network communication and 5G network communication according to various embodiments;

[0012] Figure 3 A block diagram of an example electronic device according to an embodiment is shown;

[0013] Figure 4 A block diagram of an example electronic device according to an embodiment is shown;

[0014] Figure 5 A flowchart illustrating a method for estimating the characteristics of a wireless communication channel by an electronic device according to an embodiment is shown;

[0015] Figure 6 This is a diagram illustrating an example time-slot structure of a signal received by an electronic device according to an embodiment;

[0016] Figure 7 This is a diagram illustrating a method for estimating a wireless communication channel by an electronic device according to an embodiment;

[0017] Figure 8 This is a diagram illustrating an example neural network according to an embodiment;

[0018] Figure 9This is a diagram illustrating an example neural network according to an embodiment;

[0019] Figure 10 This is a diagram illustrating an example neural network according to an embodiment;

[0020] Figure 11 This is a diagram illustrating the training cycle according to an embodiment; and

[0021] Figure 12 A flowchart illustrating a method for estimating the characteristics of a wireless communication channel by an electronic device according to an embodiment is shown. Detailed Implementation

[0022] Figure 1 This is a block diagram illustrating an electronic device 101 in a network environment 100 according to various embodiments.

[0023] refer to Figure 1 In network environment 100, electronic device 101 can communicate with electronic device 102 via a first network 198 (e.g., a short-range wireless communication network), or with at least one of electronic device 104 or server 108 via a second network 199 (e.g., a long-range wireless communication network). According to an embodiment, electronic device 101 can communicate with electronic device 104 via server 108. According to an embodiment, electronic device 101 may include a processor 120, memory 130, input module 150, sound output module 155, display module 160, audio module 170, sensor module 176, interface 177, connection terminal 178, haptic module 179, camera module 180, power management module 188, battery 189, communication module 190, user identification module (SIM) 196, or antenna module 197. In some embodiments, at least one of these components (e.g., connection terminal 178) may be omitted from electronic device 101, or one or more other components may be added to electronic device 101. In some embodiments, some of the components described above (e.g., sensor module 176, camera module 180, or antenna module 197) may be implemented as a single component (e.g., display module 160).

[0024] Processor 120 can execute, for example, software (e.g., program 140) to control at least one other component (e.g., hardware or software component) of electronic device 101 connected to processor 120, and can perform various data processing or calculations. According to one embodiment, as at least part of data processing or calculation, processor 120 can store commands or data received from another component (e.g., sensor module 176 or communication module 190) in volatile memory 132, process the commands or data stored in volatile memory 132, and store the resulting data in non-volatile memory 134. According to embodiments, processor 120 may include a main processor 121 (e.g., a central processing unit (CPU) or application processor (AP)) or an auxiliary processor 123 (e.g., a graphics processing unit (GPU), neural processing unit (NPU), image signal processor (ISP), sensor central processor, or communication processor (CP)) that is operationally independent of or combined with the main processor 121. For example, when electronic device 101 includes a main processor 121 and an auxiliary processor 123, the auxiliary processor 123 can be adapted to consume less power than the main processor 121, or adapted to be dedicated to a specific function. The auxiliary processor 123 can be implemented separately from the main processor 121, or implemented as part of the main processor 121.

[0025] When the main processor 121 is inactive (e.g., in sleep mode), the auxiliary processor 123 (rather than the main processor 121) can control at least some of the functions or states associated with at least one component of the electronic device 101 (e.g., display module 160, sensor module 176, or communication module 190), or when the main processor 121 is active (e.g., running an application), the auxiliary processor 123 can work with the main processor 121 to control at least some of the functions or states associated with at least one component of the electronic device 101 (e.g., display module 160, sensor module 176, or communication module 190). According to embodiments, the auxiliary processor 123 (e.g., an image signal processor or a communication processor) can be implemented as part of another component (e.g., camera module 180 or communication module 190) functionally associated with the auxiliary processor 123. According to embodiments, the auxiliary processor 123 (e.g., a neural processing unit) can include hardware architectures dedicated to artificial intelligence model processing. Artificial intelligence models can be generated through machine learning. For example, such learning can be performed via electronic device 101 where artificial intelligence is performed, or via a separate server (e.g., server 108). The learning algorithm can include, but is not limited to, supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning. The artificial intelligence model can include multiple layers of artificial neural networks. The artificial neural network can be a deep neural network (DNN), a convolutional neural network (CNN), a recurrent neural network (RNN), a restricted Boltzmann machine (RBM), a deep belief network (DBN), a bidirectional recurrent deep neural network (BRDNN), a deep Q-network, or a combination of two or more thereof, but is not limited thereto. Additionally or optionally, the artificial intelligence model can include software structures in addition to hardware structures.

[0026] Memory 130 may store various data used by at least one component of electronic device 101 (e.g., processor 120 or sensor module 176). For example, the various data may include input or output data of software (e.g., program 140) and commands associated therewith. Memory 130 may include volatile memory 132 or non-volatile memory 134.

[0027] The program 140 can be stored as software in the memory 130, and the program 140 may include, for example, an operating system (OS) 142, middleware 144, or application 146.

[0028] Input module 150 can receive commands or data from outside electronic device 101 (e.g., a user) that will be used by another component of electronic device 101 (e.g., processor 120). Input module 150 may include, for example, a microphone, mouse, keyboard, keys (e.g., buttons), or digital pen (e.g., stylus).

[0029] The audio output module 155 can output audio signals to the outside of the electronic device 101. The audio output module 155 may include, for example, a speaker or a receiver. The speaker can be used for general purposes such as playing multimedia or playing records. The receiver can be used to receive incoming calls. According to embodiments, the receiver can be implemented separately from the speaker, or as part of the speaker.

[0030] Display module 160 can visually provide information to the outside of electronic device 101 (e.g., to a user). Display module 160 may include, for example, a display, a holographic device, or a projector, and control circuitry for controlling a respective one of the display, holographic device, and projector. According to an embodiment, display module 160 may include a touch sensor adapted to detect touch or a pressure sensor adapted to measure the intensity of the force caused by touch.

[0031] The audio module 170 can convert sound into electrical signals and vice versa. According to an embodiment, the audio module 170 can obtain sound via the input module 150, or output sound via the sound output module 155 or headphones of an external electronic device (e.g., electronic device 102) that is directly (e.g., wired) or wirelessly connected to the electronic device 101.

[0032] Sensor module 176 can detect the operating state of electronic device 101 (e.g., power or temperature) or the environmental state outside electronic device 101 (e.g., user state), and then generate an electrical signal or data value corresponding to the detected state. According to embodiments, sensor module 176 may include, for example, a gesture sensor, gyroscope sensor, atmospheric pressure sensor, magnetic sensor, accelerometer, grip sensor, proximity sensor, color sensor, infrared (IR) sensor, biometric sensor, temperature sensor, humidity sensor, or illuminance sensor.

[0033] Interface 177 may support one or more specific protocols used to enable electronic device 101 to connect directly (e.g., wired) or wirelessly to external electronic devices (e.g., electronic device 102). According to embodiments, interface 177 may include, for example, a High Definition Multimedia Interface (HDMI), a Universal Serial Bus (USB) interface, a Secure Digital (SD) card interface, or an audio interface.

[0034] Connection end 178 may include a connector, through which electronic device 101 can be physically connected to an external electronic device (e.g., electronic device 102). According to embodiments, connection end 178 may include, for example, an HDMI connector, a USB connector, an SD card connector, or an audio connector (e.g., a headphone connector).

[0035] The haptic module 179 can convert electrical signals into mechanical stimuli (e.g., vibration or movement) or electrical stimuli that can be recognized by a user through his touch or kinesthesia. According to embodiments, the haptic module 179 may include, for example, a motor, a piezoelectric element, or an electrical stimulator.

[0036] Camera module 180 can capture still images or moving images. According to an embodiment, camera module 180 may include one or more lenses, an image sensor, an image signal processor, or a flash.

[0037] The power management module 188 can manage the power supplied to the electronic device 101. According to one embodiment, the power management module 188 may be implemented as at least part of, for example, a power management integrated circuit (PMIC).

[0038] Battery 189 can power at least one component of electronic device 101. According to embodiments, battery 189 may include, for example, a non-rechargeable primary battery, a rechargeable rechargeable battery, or a fuel cell.

[0039] Communication module 190 can support the establishment of a direct (e.g., wired) or wireless communication channel between electronic device 101 and external electronic devices (e.g., electronic device 102, electronic device 104, or server 108), and perform communication via the established communication channel. Communication module 190 may include one or more communication processors capable of operating independently of processor 120 (e.g., application processor (AP)) and support direct (e.g., wired) or wireless communication. According to embodiments, communication module 190 may include wireless communication module 192 (e.g., cellular communication module, short-range wireless communication module, or Global Navigation Satellite System (GNSS) communication module) or wired communication module 194 (e.g., local area network (LAN) communication module or power line communication (PLC) module). One of these communication modules can communicate with an external electronic device via a first network 198 (e.g., a short-range communication network such as Bluetooth™, Wi-Fi Direct, or Infrared Data Association (IrDA)) or a second network 199 (e.g., a long-range communication network such as a traditional cellular network, 5G network, next-generation communication network, the Internet, or a computer network (e.g., a LAN or a wide area network (WAN))). These various types of communication modules can be implemented as a single component (e.g., a single chip) or as multiple components separate from each other (e.g., multiple chips). The wireless communication module 192 can identify and verify the electronic device 101 in the communication network (such as the first network 198 or the second network 199) using user information (e.g., the International Mobile Subscriber Identity (IMSI)) stored in the user identification module 196.

[0040] Wireless communication module 192 can support 5G networks following 4G networks and next-generation communication technologies (e.g., new radio (NR) access technologies). NR access technologies can support enhanced mobile broadband (eMBB), massive machine-type communication (mMTC), or ultra-reliable low-latency communication (URLLC). Wireless communication module 192 can support high-frequency bands (e.g., millimeter-wave bands) to achieve, for example, high data transmission rates. Wireless communication module 192 can support various technologies used to ensure performance in high-frequency bands, such as, for example, beamforming, massive MIMO, full-dimensional MIMO (FD-MIMO), array antennas, analog beamforming, or massive antennas. Wireless communication module 192 can support various requirements specified in electronic device 101, external electronic device (e.g., electronic device 104), or network system (e.g., second network 199). According to an embodiment, the wireless communication module 192 may support peak data rates (e.g., 20 Gbps or greater) for implementing eMBB, lost coverage (e.g., 164 dB or less) for implementing mMTC, or U-plane latency (e.g., 0.5 ms or less for each of the downlink (DL) and uplink (UL), or 1 ms or less round trip) for implementing URLLC.

[0041] Antenna module 197 can transmit signals or power to or from the outside of electronic device 101 (e.g., external electronic device) or receive signals or power from the outside of electronic device 301 (e.g., external electronic device). According to an embodiment, antenna module 197 may include an antenna comprising a radiating element formed of a conductive material or conductive pattern formed in or on a substrate (e.g., a printed circuit board (PCB)). According to an embodiment, antenna module 197 may include multiple antennas (e.g., an array antenna). In this case, at least one antenna suitable for a communication scheme used in a communication network (such as a first network 198 or a second network 199) can be selected from the multiple antennas by, for example, communication module 190 (e.g., wireless communication module 192). Signals or power can then be transmitted or received between communication module 190 and external electronic device via the selected at least one antenna. According to an embodiment, additional components besides the radiating element (e.g., a radio frequency integrated circuit (RFIC)) may be additionally incorporated into antenna module 197.

[0042] According to various embodiments, antenna module 197 can form a millimeter-wave antenna module. According to embodiments, the millimeter-wave antenna module may include a printed circuit board, a radio frequency integrated circuit (RFIC), and multiple antennas (e.g., an array antenna), wherein the RFIC is disposed on or adjacent to a first surface (e.g., a bottom surface) of the printed circuit board and is capable of supporting a specified high-frequency band (e.g., a millimeter-wave band), and the multiple antennas are disposed on or adjacent to a second surface (e.g., a top or side surface) of the printed circuit board and are capable of transmitting or receiving signals in the specified high-frequency band.

[0043] At least some of the aforementioned components can be coupled to each other and transmit signals (e.g., commands or data) communicatively between them via an inter-peripheral communication scheme (e.g., bus, general purpose input / output (GPIO), serial peripheral interface (SPI), or mobile industrial processor interface (MIPI)).

[0044] According to an embodiment, commands or data can be sent or received between electronic device 101 and external electronic device 104 via server 108 connected to a second network 199. Each of electronic device 102 or electronic device 104 can be a device of the same type as electronic device 101, or a device of a different type than electronic device 101. According to an embodiment, all or some operations to be performed on electronic device 101 can be performed on one or more of external electronic devices 102, external electronic devices 104, or server 108. For example, if electronic device 101 is required to automatically perform a function or service, or should perform a function or service in response to a request from a user or another device, electronic device 101 may request one or more external electronic devices to perform at least a portion of the function or service, instead of running the function or service, or electronic device 101 may request one or more external electronic devices to perform at least a portion of the function or service in addition to running the function or service. Upon receiving the request, the one or more external electronic devices may perform the requested at least portion of the function or service, or perform additional functions or services related to the request, and transmit the result of the execution to electronic device 101. Electronic device 101 may provide the result as at least a partial response to the request, with or without further processing of the result. For this purpose, technologies such as cloud computing, distributed computing, mobile edge computing (MEC), or client-server computing can be used. Electronic device 101 may use, for example, distributed computing or mobile edge computing to provide ultra-low latency services. In another embodiment, external electronic device 104 may include Internet of Things (IoT) devices. Server 108 may be an intelligent server using machine learning and / or neural networks. According to embodiments, external electronic device 104 or server 108 may be included in a second network 199. Electronic device 101 can be applied to intelligent services (e.g., smart homes, smart cities, smart cars, or healthcare) based on 5G communication technology or IoT-related technologies.

[0045] Figure 2a This is a block diagram 200 of an electronic device 101 for supporting conventional network communication and 5G network communication according to an embodiment. (See reference...) Figure 2aThe electronic device 101 may include a first communication processor 212, a second communication processor 214, a first radio frequency integrated circuit (RFIC) 222, a second RFIC 224, a third RFIC 226, a fourth RFIC 228, a first radio frequency front-end (RFFE) 232, a second RFFE 234, a first antenna module 242, a second antenna module 244, a third antenna module 246, and an antenna 248. The electronic device 101 may also include a processor 120 and a memory 130. The second network 199 may include a first cellular network 292 and a second cellular network 294. According to another embodiment, the electronic device 101 may also include... Figure 1 At least one of the components shown, and the second network 199 may also include at least another network. According to an embodiment, the first communication processor 212, the second communication processor 214, the first RFIC 222, the second RFIC 224, the fourth RFIC 228, the first RFFE 232, and the second RFFE 234 may configure at least a portion of the wireless communication module 192. According to another embodiment, the fourth RFIC 228 may be omitted, or the fourth RFIC 228 may be included as part of the third RFIC 226.

[0046] The first communication processor 212 can establish a communication channel within the frequency band used for wireless communication with the first cellular network 292, and can support conventional network communication performed through the established communication channel. According to an embodiment, the first cellular network can be a conventional network including second-generation (2G), 3G, 4G, or Long Term Evolution (LTE) networks. The second communication processor 214 can establish a communication channel corresponding to a specified frequency band (e.g., approximately 6 GHz to 60 GHz) within the frequency band used for wireless communication with the second cellular network 294, and can support 5G network communication performed through the established communication channel. According to an embodiment, the second cellular network 294 can be a 5G network as defined in 3GPP. Additionally, according to an embodiment, the first communication processor 212 or the second communication processor 214 can establish a communication channel corresponding to another specified frequency band (e.g., approximately 6 GHz or lower) within the frequency band used for wireless communication with the second cellular network 294, and can support 5G network communication performed through the established communication channel.

[0047] The first communication processor 212 can send data to or receive data from the second communication processor 214. For example, data that has been classified as being transmitted via the second cellular network 294 can be changed to be transmitted via the first cellular network 292. In this case, the first communication processor 212 can receive the transmitted data from the second communication processor 214. For example, the first communication processor 212 can send data to or receive data from the second communication processor 214 via an inter-processor interface 213. The inter-processor interface 213 can be implemented as, for example, a Universal Asynchronous Receiver / Transmitter (UART) (e.g., High-Speed ​​UART (HS-UART)) or a Peripheral Component Interconnect Bus Fast (PCIe) interface, but its type is not limited. Alternatively, the first communication processor 212 and the second communication processor 214 can exchange control information and packet data information by using, for example, shared memory. The first communication processor 212 can send various information to or receive various information from the second communication processor 214, such as sensing information, information about output strength, and resource block (RB) allocation information.

[0048] According to an implementation, the first communication processor 212 may not be directly connected to the second communication processor 214. In this case, the first communication processor 212 can send data to or receive data from the second communication processor 214 via the processor 120 (e.g., an application processor). For example, the first communication processor 212 and the second communication processor 214 can send data to or receive data from each other via the processor 120 (e.g., an application processor) and an HS-UART interface or a PCIe interface, but the type of interface is not limited. Alternatively, the first communication processor 212 and the second communication processor 214 can exchange control information and packet data information by using the processor 120 (e.g., an application processor) and shared memory.

[0049] According to embodiments, the first communication processor 212 and the second communication processor 214 can be implemented in a single chip or a single package. According to embodiments, the first communication processor 212 or the second communication processor 214 can be configured together with the processor 120, the auxiliary processor 123, or the communication module 190 in a single chip or a single package. For example, as... Figure 2b As shown, the integrated communication processor 260 can support both functions for communicating with the first cellular network 292 and functions for communicating with the second cellular network 294.

[0050] As described above, at least one of processor 120, first communication processor 212, second communication processor 214, or integrated communication processor 260 can be implemented as a single chip or a single package. In this case, the single chip or single package may include: a memory (or storage device) storing instructions that cause at least some operations to be performed according to various embodiments; and processing circuitry (or, without limitation, such as computing circuitry) for executing the instructions. The instructions stored in the memory, when executed individually or jointly by at least one processor, can cause electronic device 101 to perform at least one operation.

[0051] The first RFIC 222 can, during transmission, convert the baseband signal generated by the first communication processor 212 into a radio frequency (RF) signal with a frequency of approximately 700 MHz to approximately 3 GHz, which is used in the first cellular network 292 (e.g., a conventional network). During reception, the RF signal can be obtained from the first network 292 (e.g., a conventional network) via an antenna (e.g., the first antenna module 242) and can be preprocessed by an RFFE (e.g., the first RFFE 232). The first RFIC 222 can convert the preprocessed RF signal back into a baseband signal so that the preprocessed RF signal can be processed by the first communication processor 212.

[0052] The second RFIC 224 can, during transmission, convert the baseband signal generated by the first communication processor 212 or the second communication processor 214 into an RF signal (hereinafter referred to as a 5G Sub6 RF signal) within the Sub6 frequency band (e.g., approximately 6 GHz or lower) used in the second cellular network 294 (e.g., a 5G network). During reception, the 5G Sub6 RF signal can be obtained from the second cellular network 294 (e.g., the 5G network) via an antenna (e.g., the second antenna module 244) and can be preprocessed via an RFFE (e.g., the second RFFE 234). The second RFIC 224 can convert the preprocessed 5G Sub6 RF signal back into a baseband signal so that the preprocessed 5G Sub6 RF signal can be processed by the corresponding communication processor in the first communication processor 212 or the second communication processor 214.

[0053] The third RFIC 226 can convert the baseband signal generated by the second communication processor 214 into an RF signal (hereinafter referred to as the 5G Above6 RF signal) within the 5G Above6 frequency band (e.g., about 6 GHz to about 60 GHz) to be used in the second cellular network 294 (e.g., a 5G network). Upon reception, the 5G Above6 RF signal can be obtained from the second cellular network 294 (e.g., the 5G network) via an antenna (e.g., antenna 248) and can be preprocessed via the third RFFE 236. The third RFIC 226 can convert the preprocessed 5G Above6 RF signal back into a baseband signal so that the preprocessed 5G Above6 RF signal can be processed by the second communication processor 214. According to an embodiment, the third RFFE 236 can be configured as part of the third RFIC 226.

[0054] According to an embodiment, the electronic device 101 may include a fourth RFIC 228, separate from or at least part of the third RFIC 226. In this case, the fourth RFIC 228 may convert the baseband signal generated by the second communication processor 214 into an RF signal (hereinafter referred to as an IF signal) in the intermediate frequency band (e.g., about 9 GHz to 11 GHz), and then transmit the IF signal to the third RFIC 226. The third RFIC 226 may convert the IF signal into a 5G Above6 RF signal. Upon reception, the 5G Above6 RF signal may be received from the second cellular network 294 (e.g., a 5G network) via an antenna (e.g., antenna 248) and may be converted into an IF signal by the third RFIC 226. The fourth RFIC 228 may convert the IF signal back into a baseband signal so that the IF signal can be processed by the second communication processor 214.

[0055] According to embodiments, the first RFIC 222 and the second RFIC 224 can be implemented as at least a portion of a single chip or a single package. According to embodiments, such as Figure 2a or Figure 2bAs shown, when the first RFIC 222 and the second RFIC 224 are implemented as a single chip or a single package, the first RFIC and the second RFIC can be implemented as an integrated RFIC. In this case, the integrated RFIC can be connected to the first RFFE 232 and the second RFFE 234 to convert the baseband signal into a signal within the frequency band supported by the first RFFE 232 and / or the second RFFE 234, and send the converted signal to one of the first RFFE 232 and the second RFFE 234. According to an embodiment, the first RFFE 232 and the second RFFE 234 can be implemented as at least part of a single chip or a single package. According to an embodiment, at least one of the first antenna module 242 or the second antenna module 244 can be omitted or combined with another antenna module to process RF signals within multiple corresponding frequency bands.

[0056] According to an embodiment, the third RFIC 226 and antenna 248 can be arranged on the same substrate to configure the third antenna module 246. For example, the wireless communication module 192 or processor 120 can be disposed on the first substrate (e.g., the main PCB). In this case, the third RFIC 226 can be disposed in a portion of a second substrate (e.g., a sub-PCB) separate from the first substrate (e.g., the lower surface), and the antenna 248 can be disposed in another portion of the substrate (e.g., the upper surface), thereby configuring the third antenna module 246. Arranging the third RFIC 226 and antenna 248 on the same substrate can shorten the length of the transmission line between them. This can, for example, reduce the loss (e.g., attenuation) of the transmission line to signals in the high-frequency band (e.g., from about 6 GHz to about 60 GHz) used in 5G network communication. Therefore, the electronic device 101 can improve the quality or speed of communication with the second network 294 (e.g., a 5G network).

[0057] According to an embodiment, antenna 248 may be configured as an antenna array comprising a plurality of antenna elements that can be used for beamforming. In this case, the third RFIC 226 may include, for example, a plurality of phase shifters 238 corresponding to the plurality of antenna elements as part of the third RFFE 236. During transmission, each of the plurality of phase shifters 238 may shift the phase of a 5G Above6 RF signal to be transmitted through the corresponding antenna element to an external location (e.g., a base station of a 5G network) of electronic device 101. During reception, each of the plurality of phase shifters 238 may shift the phase of a 5G Above6 RF signal already received from the external location through the corresponding antenna element to the same or substantially the same phase. This enables beamforming transmission or reception between electronic device 101 and an external location.

[0058] The second cellular network 294 (e.g., a 5G network) can operate independently of the first cellular network 292 (e.g., a legacy network) (e.g., standalone operation (SA)), or it can operate while connected to the first cellular network 292 (e.g., non-standalone operation (NSA)). For example, a 5G network may only have an access network (e.g., a 5G radio access network (RAN) or a next-generation RAN (NG RAN)) without a core network (e.g., a next-generation core (NGC)). In this case, electronic device 101 can access the access network of the 5G network and then access an external network (e.g., the Internet) under the control of the core network of the legacy network (e.g., an evolved packet core (EPC)). Protocol information for communicating with the legacy network (e.g., LTE protocol information) or protocol information for communicating with the 5G network (e.g., New Radio (NR) protocol information) is stored in memory 230 and can be accessed by another component (e.g., processor 120, a first communication processor 212, or a second communication processor 214).

[0059] Figure 3 A block diagram of an example electronic device (e.g., electronic device 101) according to an embodiment is shown.

[0060] In an embodiment, electronic device 101 may include a plurality of antennas 310, memory 320 and / or processor 330.

[0061] In an embodiment, electronic device 101 may receive signals from an external electronic device via at least one of its multiple antennas 310. If electronic device 101 is implemented as at least part of a base station, the external electronic device may be a terminal. If electronic device 101 is implemented as a portable electronic device (e.g., a smartphone, wearable device, or tablet computer), the external electronic device may be a base station.

[0062] In an embodiment, memory 320 may include an artificial neural network for estimating the characteristics of a wireless communication channel. Memory 320 may also include an artificial neural network for estimating the wireless communication channel (e.g., calculating a distribution associated with the type of wireless communication channel or filter coefficients of the wireless communication channel). The artificial neural network can be trained using field data obtained based on received signals. Memory 320 may store training data used to train the artificial neural network. (Refer to...) Figure 4 A detailed method is described for obtaining training data (e.g., ground truth) for training an artificial neural network by electronic device 101.

[0063] In an embodiment, processor 330 may perform general operations for training an artificial neural network for estimating the characteristics of a wireless communication channel. Processor 330 may include processing circuitry (not shown) for performing the operations described below. In this disclosure, operations performed by a specific module (or block) can be understood as being performed by processor 330.

[0064] Figure 4 It is a block diagram illustrating the configuration of an electronic device.

[0065] refer to Figure 4 In an embodiment, the electronic device (e.g., electronic device 101) may include a Fast Fourier Transform (FFT) block 410, a channel estimation block 420, a neural network block 430, an equalizer 440, and a decoder 450. Each of the FFT block 410, channel estimation block 420, neural network block 430, equalizer 440, and decoder 450 may be implemented as an independent circuit or as part of a receiver. The operations performed by each of the FFT block 410, channel estimation block 420, neural network block 430, equalizer 440, and decoder 450 can be understood as operations performed by a processor (e.g., processor 330).

[0066] In an embodiment, FFT block 410 can transform the received signal based on FFT. The signal transformed by FFT block 410 can be transmitted to channel estimation block 420 or equalizer 440. For example, the frequency domain signal (e.g., Y_RS) transformed by FFT block 410 can be transmitted to channel estimation block 420. For example, the data signal (e.g., Y_data) transformed by FFT block 410 can be transmitted to equalizer 440. FFT block 410 can, for example, transform the received signal corresponding to the nth antenna and the i-th RS symbol into a frequency domain signal based on Equation 1. The “wireless communication channel” can be an uplink channel for transmitting signals from a terminal to a base station, but is not limited to this. For example, the wireless communication channel can be a downlink channel for transmitting signals from a base station to a terminal. The signal received by the electronic device can be an RF signal as defined in the New Radio (NR) specification, but is not limited to this. For example, the received signal can be a signal defined in the Long Term Evolution (LTE) or 6th Generation (6G) specifications.

[0067] [Equation 1]

[0068]

[0069] In an embodiment, It can be the channel matrix of the k-th subcarrier. It can be the pilot sequence of the k-th subcarrier. It can be additive white Gaussian noise (AWGN) of the k-th subcarrier. Here, k can be such as 0, 1, or... Natural numbers such as -1. On the reference symbol (“RS”), a pilot sequence previously defined between the base station and the terminal can be transmitted. Each resource block may include six pilot tones. It corresponds to 6 times the allocated resource block (Allocated_RB) and can be the number of pilot tones allocated to the reference symbol.

[0070] In an embodiment, channel estimation block 420 can perform real-time estimation of information associated with the characteristics of the wireless communication channel on a time-slot basis. Characteristics of the wireless communication channel may include, for example, power delay distribution (PDP) or Doppler spread (DOP), and are not limited thereto. Characteristics of the wireless communication channel may include characteristics of fading channels, such as time offset, frequency offset, LOS, non-line-of-sight (NLOS), and the angle of arrival of the signal. Estimation of the wireless communication channel between the terminal (user equipment) and the base station can be an important factor in improving the reception performance of a mobile communication system. PDP can represent the time delay value of reflected waves in the fading channel environment between the terminal and the base station, as well as the power of each reflected wave component. PDP can have consistent characteristics on a cell-by-cell basis provided by the base station. For example, in an urban environment, various reflected waves can be received. In a non-urban environment, signals with direct wave (line-of-sight "LOS") components can be received. The channel in a non-urban environment can be a flat fading channel with a small number of multipaths. Estimation of PDP can help improve complexity in the frequency domain and channel estimation performance. For example, electronic device 101 can adaptively optimize the filter coefficients used for channel estimation based on the estimated PDP. Electronic device 101 can configure the channel estimation filter coefficients corresponding to TDL-A as filter coefficients based on the identification that the fading channel of the cell is TDL-A. Doppler spread may occur due to the relative speed of the terminal relative to the base station depending on the terminal's movement. Doppler spread can be a channel characteristic that affects the time-axis fading variance of the fading channel between the terminal and the base station. Doppler spread can also have consistent characteristics on a cell-by-cell basis. For example, high Doppler spread can be detected in cells covering environments surrounding highways or high-speed railways. Estimating Doppler spread can help improve the complexity in the time domain and the performance of channel estimation.

[0071] In an embodiment, the channel estimation block 420 may include a wireless channel characteristic acquisition unit 421 and a channel estimation neural network 423. The wireless channel characteristic acquisition unit 421 may output information associated with the instantaneous characteristics of the wireless communication channel. For example, the wireless channel characteristic acquisition unit 421 may output an instantaneous PDP (PDP_inst). The wireless channel characteristic acquisition unit 421 may output an instantaneous DOP (DOP_inst). The wireless channel characteristic acquisition unit 421 may output both instantaneous PDP and instantaneous DOP. The wireless channel characteristic acquisition unit 421 may estimate the instantaneous PDP and / or instantaneous DOP by using a method based on a non-artificial neural network algorithm (e.g., non-AI). Algorithms that do not employ artificial neural networks may include signal processing techniques. The wireless channel characteristic acquisition unit 421 may also estimate the instantaneous PDP and / or instantaneous DOP by using a general neural network model based on offline training. The wireless channel characteristic acquisition unit 421 may decorrelate the pilot sequence of the received signal in the frequency domain based on Equation 2 to obtain the wireless channel characteristics.

[0072] [Equation 2]

[0073]

[0074] According to Equation 2, the relevant reference symbols can be removed from Equation 3. Vector representation of pilot tones.

[0075] [Equation 3]

[0076]

[0077] The wireless channel characteristic acquisition unit 421 can be based on Equation 4. Perform the inverse discrete Fourier transform (IDFT) to obtain the signal in the time domain.

[0078] [Equation 4]

[0079]

[0080] The wireless channel characteristic acquisition unit 421 can estimate the instantaneous PDP of the nth antenna and the ith RS symbol based on Equation 5, using time slots as units, according to the signal in the time domain.

[0081] [Equation 5]

[0082]

[0083] It can be a function for calculating instantaneous PDP, and can be implemented using neural networks or non-AI algorithms trained offline.

[0084] The wireless channel characteristics acquisition unit 421 can estimate the instantaneous DOP based on Equation 6.

[0085] [Equation 6]

[0086]

[0087] It can be an instantaneous Doppler extension estimated for a specific time slot. The function () can be used to calculate the instantaneous Doppler spread and can be implemented using a neural network or a non-AI algorithm based on offline training. The instantaneous PDP and instantaneous DOP obtained by the channel characteristic acquisition unit 421 can be provided to the neural network block 430 as input for online training. The channel estimation neural network 423 can obtain the estimated instantaneous channel value H_est based on the decorrelated reference symbols and the instantaneous PDP and instantaneous DOP obtained by the wireless channel characteristic acquisition unit 421. The estimated instantaneous channel value can be transmitted to the equalizer 440. The channel estimation result and the data signal (Y_DATA) output by the FFT block 410 can be input to the decoder 450 via the equalizer 440. The decoder 450 can decode the input signal. Information associated with the packets decoded by the decoder 450 can be provided to the neural network block 430. The decoder 450 can provide the neural network block 430 with, for example, the cyclic redundancy check (CRC) result value of the packets decoded from the received signal.

[0088] In an embodiment, the neural network block 430 may include a reliability determination unit 431, a true label generation unit 433, and an artificial neural network 435. To perform online training, reliable field data may need to be collected. The reliability determination unit 431 may identify the reliability of the instantaneously estimated PDP (PDP_inst) based on the CRC results of packets decoded from the received signal. For example, if the CRC result of a packet decoded from the received signal corresponds to a value indicating that there are no errors in the decoded packet (e.g., an OK response), the reliability determination unit 431 may identify the instantaneously estimated channel characteristics (e.g., PDP_inst and / or DOP_inst) corresponding to the error-free time slot as true. If the CRC result of a packet decoded from the received signal corresponds to a value indicating that there are errors in the decoded packet (e.g., an error response), the reliability determination unit 431 may identify the estimated signal-to-noise ratio (SNR) on a time slot basis. If the SNR of the received signal exceeds a threshold, the reliability determination unit 431 may determine the instantaneously estimated channel characteristics as reliable training data even if the CRC result of the decoded packet is not OK. In an embodiment, the reliability determination unit 431 can determine the instantaneously estimated channel characteristics as reliable training data only if the CRC result of the packet decoded from the received signal corresponds to a value indicating that there are no errors in the decoded packet and the SNR of the received signal exceeds a threshold. In an embodiment, the reliability determination unit 431 can use Long Delay Response (LLR) or SNR to identify the reliability of the instantaneously estimated channel characteristics for channels without channel codecs (e.g., Sounding Reference Signal (SRS) or Physical Random Access Channel (PRACH)). The reliability determination unit 431 can also identify the reliability of the instantaneously estimated channel characteristics by recognizing the CRC result for channels with channel codecs (e.g., Physical Uplink Shared Channel (PUSCH) or Physical Uplink Control Channel (PUCCH)). The real tag generation unit 433 can generate real tags based on the reliability identification result of the reliability determination unit 431. For example, if the reliability determination unit 431 determines the instantaneously estimated channel characteristics as reliable data, the real tag generation unit 433 can store the received signal and the instantaneously estimated channel characteristics as a real tag pair. If the reliability determination unit 431 determines that the instantaneously estimated channel characteristics are unreliable data, the true label generation unit 433 may not store the instantaneously estimated channel characteristics corresponding to the time slot. The true label generation unit 433 may, for example, discard unreliable packets. The true label generation unit 433 may provide true label pairs to the artificial neural network 435. The artificial neural network 435 may perform online training based on learning the provided true label pairs.The artificial neural network 435 can perform online training on the characteristics of the received signal on a cell-by-cell basis, thereby improving the estimation performance of the wireless channel characteristics corresponding to the cell. In an embodiment, the artificial neural network 435 can perform online training for channel estimation and training for channel characteristics. For example, based on the estimation results of the wireless communication channel (e.g., the filter coefficients of the wireless communication channel or the type of the wireless communication channel) satisfying the conditions of the online training data, the artificial neural network 435 can store the estimation results of the wireless communication channel as true values. After performing online training, the artificial neural network 435 can estimate the characteristics of the wireless channel based on the reference symbols of the received signal transformed by the FFT block 410. Compared with wireless channel characteristics estimated by a neural network model based on offline training or non-AI estimation, the wireless channel characteristics estimated by the artificial neural network 435 based on cell-specific online training can have improved estimation accuracy.

[0089] Figure 5 This is a flowchart illustrating a method for online training of an artificial neural network in an electronic device. (Refer to...) Figure 6 , Figure 7 , Figure 8 , Figure 9 , Figure 10 and Figure 11 describe Figure 5 Examples of implementations. Figure 6 This is a diagram of an example time slot structure of a signal received by an electronic device according to an embodiment. Figure 7 This is a diagram illustrating a method for estimating a wireless communication channel by an electronic device according to an embodiment. Figure 8 This is a diagram illustrating an example neural network according to an embodiment. Figure 9 This is a diagram illustrating an example neural network according to an embodiment. Figure 10 This is a diagram illustrating an example neural network according to an embodiment. Figure 11 This is a diagram illustrating the training cycle according to an embodiment.

[0090] refer to Figure 5 In this embodiment, electronic device 101 (e.g., processor 330) may receive signals in operation 501. Electronic device 101 may receive signals from an external electronic device via at least one of its multiple antennas. If electronic device 101 is implemented as at least part of a base station, the external electronic device may be a terminal. If electronic device 101 is implemented as a portable electronic device (e.g., a smartphone, wearable device, or tablet computer), the external electronic device may be a base station.

[0091] In an embodiment, during operation 503, electronic device 101 can identify whether information associated with characteristics of the wireless communication channel (or information associated with estimation results of the wireless communication channel) identified from the received signal satisfies the conditions for identifying the reliability of the training data. Electronic device 101 can identify information associated with characteristics of the wireless communication channel on a time-slot basis based on the received signal. Electronic device 101 can estimate the wireless communication channel itself, such as the distribution corresponding to the type of wireless communication channel or the filter coefficients of the wireless communication channel. (See reference...) Figure 6 On reference symbols 611 and 613, a pilot sequence previously defined between the base station and the terminal can be transmitted. For example, the number of pilot tones corresponding to a resource block can be 6, and there is no specific limit to the number of pilot tones.

[0092] In embodiments, information associated with the characteristics of the wireless communication channel may include, for example, instantaneous PDP and instantaneous Doppler spread, and examples of characteristics of the wireless communication channel are not limited thereto. Electronic device 101 may input the received signal into an artificial neural network that has not undergone online training to obtain the instantaneous PDP and instantaneous Doppler spread based on information from the output of the artificial neural network that has not undergone online training. The artificial neural network that has not undergone online training may be a neural network based on offline training. Electronic device 101 may also obtain the instantaneous PDP and instantaneous Doppler spread based on non-AI-based signal processing. To identify whether the data collected in real time is sufficiently reliable for online training, electronic device 101 may identify whether the information associated with the characteristics of the wireless communication channel meets the conditions used to identify the reliability of the training data. Reference Figure 7 Electronic device 101 (e.g., channel estimation neural network 423) can be based on decorrelated reference symbols ( To obtain the estimated instantaneous channel values ​​corresponding to the instantaneous PDP and instantaneous DOP ( ).

[0093] In an embodiment, the condition for identifying the reliability of the training data may include a condition that the CRC result of the packets decoded from the received signal corresponds to a value indicating that there are no errors in the decoded packets. Electronic device 101 can identify whether the CRC result of the decoded packets corresponds to a value indicating that there are no errors in the decoded packets. Based on the identification that the CRC result corresponds to a value indicating that there are no errors in the decoded packets, electronic device 101 can identify that the instantaneous PDP and instantaneous Doppler spread corresponding to the time slot are reliable data, and use them as training data for online training.

[0094] In an embodiment, the conditions for identifying the reliability of the training data may include a condition that the CRC result of the packets decoded from the received signal corresponds to a value indicating the presence of an error in the decoded packets and that the SNR of the received signal exceeds a threshold. Electronic device 101 may identify whether the SNR of the received signal exceeds the threshold based on the identification that the CRC result does not correspond to a value indicating the absence of an error in the decoded packets. Electronic device 101 may identify that the instantaneous PDP and instantaneous Doppler spread corresponding to the time slot are reliable data, and use them as training data for online training, based on the identification that the SNR of the received signal exceeds the threshold.

[0095] In an embodiment, the conditions for identifying the reliability of the training data may include the condition that the CRC result of the decoded packets from the received signal corresponds to a value indicating that there are no errors in the decoded packets and that the SNR of the received signal exceeds a threshold. Electronic device 101 may identify the instantaneous PDP and instantaneous Doppler spread corresponding to the time slot as reliable data, based on the identification that the CRC result of the decoded packets corresponds to a value indicating that there are no errors in the decoded packets and that the SNR of the received signal exceeds the threshold, and use this as training data for online training.

[0096] In an embodiment, based on the recognition that information associated with the characteristics of the wireless communication channel identified from the received signal satisfies the condition for recognizing the reliability of the training data (operation 503 - Yes), electronic device 101 can perform online training of an artificial neural network for estimating the characteristics of the wireless communication channel in operation 505. Electronic device 101 can perform online training of the artificial neural network based on providing training data determined to be true labels as input to the artificial neural network. In an embodiment, electronic device 101 can determine the information associated with the estimation result of the wireless communication channel as the true value of the training data for training the artificial neural network based on the recognition that the information associated with the characteristics of the wireless communication channel satisfies the condition. Electronic device 101 can determine the reliability of real-time collected field data to provide training data for effectively performing online training of the artificial neural network.

[0097] refer to Figure 8 In this embodiment, the artificial neural network can be implemented as a generative neural network. The artificial neural network may include, for example, an architecture of a generative adversarial network (“GAN”). The artificial neural network may include, for example, a generator 810 and a discriminator 820.

[0098] In an embodiment, information associated with channel characteristics (or information associated with channel estimation results) (which serves as input to the truth value) can be input to discriminator 820. Electronic device 101 can perform identification 431 based on information associated with packets decoded by decoder 450 and information associated with the characteristics of the wireless communication channel estimated in real time by channel estimation block 420, to determine whether the CRC result corresponds to a value indicating the absence of errors in the decoded packets (e.g., an OK response). The information associated with the characteristics of the wireless communication channel estimated in real time may include, but is not limited to, instantaneous PDP. For example, the information associated with the characteristics of the wireless communication channel estimated in real time may also include instantaneous Doppler spread. The information associated with the characteristics of the wireless communication channel estimated in real time may also include both instantaneous PDP and instantaneous Doppler spread. The information associated with the characteristics of the wireless communication channel may also include at least one of various parameters indicating the characteristics of the fading channel. Electronic device 101 can provide discriminator 820 with information associated with channel characteristics based on the identification that no errors exist in the decoded packets. Discriminator 820 can distinguish between channel characteristics (or channel estimation results) input as true values ​​and channel characteristics generated by the generator, so as to be trained to identify channel characteristics input as true data and channel characteristics generated by generator 810 as fake data. Generator 810 can generate channel characteristics similar to channel characteristics input as true values, so that discriminator 820 identifies channel characteristics generated by generator 810 as true data. Artificial neural networks can be implemented as neural networks that learn the probability distribution of learning data and generative adversarial networks. Artificial neural networks can also include architectures such as variational autoencoders or diffusion models.

[0099] refer to Figure 9 In an embodiment, the artificial neural network 435 may be implemented as at least a portion of the analytical neural network 920.

[0100] In an embodiment, information associated with channel characteristics as a true value input (or information associated with channel estimation results) can be input to the analysis neural network 920. Electronic device 101 can perform an identification of whether a CRC result corresponds to a value indicating the absence of errors in the decoded packets (e.g., an OK response) based on information associated with packets decoded by decoder 450 and information associated with characteristics of the wireless communication channel estimated in real time by channel estimation block 420. Electronic device 101 can provide information associated with channel characteristics as a true label to analysis neural network 920 based on the identification that no errors exist in the decoded packets. Electronic device 101 can provide information associated with channel characteristics as a true label to analysis neural network 920 based on the absence of errors in the CRC result and an SNR exceeding a threshold SNR. Analysis neural network 920 can use the result of re-estimating channel characteristics as a true label based on data from re-estimation block 910. Analysis neural network 920 can calculate a loss function 921 based on comparing the channel estimation result (e.g., PDP) with the true label. The channel estimation result may include information associated with characteristics such as Doppler spread. The analysis of neural network 920 can be based on the loss function 921, and backpropagation 923 can be performed.

[0101] refer to Figure 10 In an embodiment, the artificial neural network 435 used for online training on a cell-by-cell basis can perform online training for two or three cases distinguished by SNR.

[0102] In this embodiment, if the SNR exceeds a first threshold, the electronic device 101 can identify that it is located in a strong electric field. Based on the SNR exceeding the first threshold, a neural network for strong electric fields can be trained online for time slots identified as having no CRC result errors.

[0103] In this embodiment, if the SNR exceeds a second threshold which is smaller than a first threshold, the electronic device 101 can identify that it is located in a moderate electric field. Based on the SNR exceeding the second threshold but less than the first threshold, a neural network for moderate electric fields can be trained online for time slots identified as having no CRC result errors.

[0104] In this embodiment, if the SNR is less than a second threshold, the electronic device 101 can detect that it is located in a weak electric field. Based on the SNR being less than the second threshold, a neural network for weak electric fields can be trained online for time slots identified as having no CRC result errors.

[0105] In one embodiment, during operation 507, electronic device 101 may input the received signal into an online-trained artificial neural network to estimate the characteristics of the wireless communication channel based on information from the output of the online-trained artificial neural network. Electronic device 101 may update the neural network using cell-specific parameters or terminal-specific parameters, based on the completion of online training. The online-trained parameters can be used to update the cell-specific neural network to provide channel characteristic estimation performance with higher accuracy than that estimated using an offline-trained neural network or a non-AI signal processing method.

[0106] In an embodiment, electronic device 101 can identify whether conditions for updating the parameters of an artificial neural network are met based on estimating channel characteristics using an online-trained neural network. Conditions for updating the parameters of the artificial neural network may include a condition that the quality of the received signal is below a threshold. The quality of the received signal may be a CRC result and / or a signal-to-noise ratio (SNR). Electronic device 101 can, for example, when estimating channel characteristics using an online-trained neural network, identify that conditions for updating the parameters of the artificial neural network are met based on identifying packets in which CRC results are incorrect. Electronic device 101 can also identify that conditions for updating the parameters of the artificial neural network are met when estimating channel characteristics using an online-trained neural network, based on identifying that the SNR is less than a threshold used to determine whether fine-tuning should be performed. Electronic device 101 can update the parameters to perform fine-tuning of the artificial neural network based on the identification that a neural network retraining condition (parameter retraining condition "PRC") is met.

[0107] refer to Figure 11In an embodiment, electronic device 101 (e.g., a portable electronic device that transmits or receives signals from a base station) can perform fine-tuning of the artificial neural network based on a period 1110 for updating the parameters of the artificial neural network. Electronic device 101 can perform online training of the artificial neural network on a UE-by-UE basis. Electronic device 101 can perform online training of the artificial neural network for estimating the characteristics of the downlink channel within at least a portion of a time interval 1111 of the period 1110 for updating the parameters of the artificial neural network. After performing online training of the artificial neural network, electronic device 101 can estimate the characteristics of the wireless communication channel using the artificial neural network within a time interval 1113. Based on the past of the period 1110 for updating the parameters of the artificial neural network, electronic device 101 can perform online training of the artificial neural network for estimating the characteristics of the downlink channel within at least a portion of a time interval 1121 of the next period. During the time interval 1121 of performing online training of the artificial neural network, electronic device 101 can stop the operation of estimating the characteristics of the wireless communication channel using the artificial neural network. After performing online training of an artificial neural network, electronic device 101 can estimate the characteristics of the wireless communication channel within a time interval 1123 using the artificial neural network. Electronic device 101 can periodically update the parameters of the artificial neural network based on its ability to move between different cells. The period 1111 for updating the parameters of the artificial neural network can be configured to correspond to information associated with the moving speed of electronic device 101. For example, the greater the moving speed of electronic device 101, the shorter the training period 1111 of the artificial neural network can be configured. Conversely, the smaller the moving speed of electronic device 101, the longer the training period 1111 of the artificial neural network can be configured.

[0108] Figure 12 A flowchart illustrating a method for estimating the characteristics of a wireless communication channel by an electronic device according to an embodiment is shown.

[0109] refer to Figure 12 In one embodiment, the electronic device 101 (e.g., processor 330) can perform online training of an artificial neural network.

[0110] In an embodiment, during operation 1201, an Online Training Flag (OTF) corresponding to the artificial neural network can be identified. The OTF corresponding to the artificial neural network can include, for example, 0, 1, and 2, and there are no limitations on specific examples of the OTF corresponding to the artificial neural network. If the OTF corresponding to the artificial neural network is 0, the characteristics of the wireless communication channel can be estimated by a general artificial neural network that does not perform online training. If the OTF corresponding to the artificial neural network is 0, the characteristics of the wireless communication channel can also be estimated based on a non-AI algorithm. If the OTF corresponding to the artificial neural network is 1, the artificial neural network can perform online training. If the OTF corresponding to the artificial neural network is 2, the characteristics of the wireless communication channel can be estimated based on an artificial neural network trained online.

[0111] In an embodiment, based on the identification that the OTF corresponding to the artificial neural network is 0, in operation 1203, electronic device 101 can estimate the characteristics of the wireless communication channel. Electronic device 101 can estimate the characteristics of the wireless communication channel by using, for example, a general artificial neural network that does not perform online training. Electronic device 101 can also estimate the characteristics of the wireless communication channel based on non-AI algorithms.

[0112] In one embodiment, during operation 1205, electronic device 101 may perform decoding based on signals received through at least a portion of a plurality of antennas. Electronic device 101 may obtain decoded packets based on the received signals via an equalizer and a decoder.

[0113] In an embodiment, during operation 1207, electronic device 101 can identify whether the decoded packet meets the conditions used to determine reliability. Electronic device 101 can identify, for example, whether the CRC result of the decoded packet corresponds to a value indicating that there are no errors in the decoded packet. Electronic device 101 can also identify whether the SNR of the received signal exceeds a threshold SNR.

[0114] In an embodiment, based on the identification that the decoded packets meet the conditions for determining reliability (operation 1207 - Yes), in operation 1209, electronic device 101 may update the OTF corresponding to the artificial neural network. Electronic device 101 may identify that the estimated channel characteristics corresponding to the corresponding time slot can be used for online training, based on the identification that the decoded packets meet the conditions for determining reliability. Electronic device 101 may, for example, update the OTF corresponding to the artificial neural network from 0 to 1.

[0115] In this embodiment, based on the identification that the OTF corresponding to the artificial neural network is 1, in operation 1211, the electronic device 101 can update the parameters of the artificial neural network. The electronic device 101 can perform online training of the artificial neural network based on channel characteristic estimates determined to be reliable. Once online training is complete, the electronic device 101 can update the artificial neural network using parameters specifically trained for the cell or terminal.

[0116] In this embodiment, the parameters based on the artificial neural network are updated, and in operation 1213, the electronic device 101 can update the OTF. For example, the electronic device 101 can update the OTF corresponding to the artificial neural network from 1 to 2.

[0117] In one embodiment, based on the identification that the OTF corresponding to the artificial neural network is 2, in operation 1215, the electronic device 101 can estimate the characteristics of the wireless communication channel. In another embodiment, in operation 1217, the electronic device 101 can perform decoding based on signals received through at least a portion of a plurality of antennas.

[0118] In one embodiment, in operation 1219, electronic device 101 can monitor the quality of the received signal. Electronic device 101 can identify whether the quality of the received signal meets the conditions (or retraining conditions, "PRC") for performing fine-tuning of the artificial neural network. In one embodiment, in operation 1221, electronic device 101 can update the OTF. Electronic device 101 can update the OTF from 2 to 0 based on the identification that the quality of the received signal is below a threshold quality.

[0119] According to an embodiment, an electronic device may include a plurality of antennas. The electronic device may include at least one processor. The electronic device may include a memory configured to store at least one instruction. The at least one processor may be configured to receive signals from an external electronic device via at least one of the plurality of antennas. The at least one processor may be configured to: identify whether information associated with characteristics of a wireless communication channel identified from the received signals satisfies a condition for identifying the reliability of training data. The at least one processor may be configured to: perform online training of an artificial neural network for estimating the characteristics of a wireless communication channel corresponding to a cell, using the information associated with the characteristics of the wireless communication channel, based on the condition that the identified information associated with the characteristics of the wireless communication channel satisfies the condition. The at least one processor may be configured to: estimate the characteristics of the wireless communication channel based on information output from the online-trained artificial neural network by inputting signals received via at least a portion of the plurality of antennas into the online-trained artificial neural network.

[0120] In an embodiment, information associated with the characteristics of the wireless communication channel may include instantaneous power delay distribution (PDP) and instantaneous Doppler spread.

[0121] In an embodiment, at least one processor may also be configured to: obtain the instantaneous PDP and instantaneous Doppler extension based on information from the output of an artificial neural network that has never undergone online training, by inputting the received signal into the artificial neural network that has not undergone online training.

[0122] In an embodiment, the conditions for identifying the reliability of the training data may include a cyclic redundancy check (CRC) result of the packets decoded from the received signal corresponding to a value indicating that there are no errors in the decoded packets.

[0123] In an embodiment, the conditions for identifying the reliability of training data may include a condition where the CRC result of the packets decoded from the received signal corresponds to a value indicating an error in the decoded packets and the signal-to-noise ratio (SNR) of the received signal exceeds a threshold.

[0124] In an embodiment, the conditions for identifying the reliability of the training data may include the condition that the CRC result of the packet decoded from the received signal corresponds to a value indicating that there are no errors in the decoded packet and that the SNR of the received signal exceeds a threshold.

[0125] In an embodiment, at least one processor may also be configured to: determine the information associated with the estimation result of the wireless communication channel as the true value of the training data for training the artificial neural network based on the recognition that the information associated with the estimation result of the wireless communication channel satisfies the condition.

[0126] In an embodiment, at least one processor may also be configured to identify whether a condition for updating the parameters of the artificial neural network is met. At least one processor may also be configured to perform fine-tuning of the artificial neural network by updating the parameters based on the identification that the condition is met.

[0127] In an embodiment, the conditions for updating the parameters of the artificial neural network may include the condition that the quality of the received signal is below a threshold.

[0128] In an embodiment, at least one processor may also be configured to perform fine-tuning of the artificial neural network based on the period used to update the parameters of the artificial neural network.

[0129] In one embodiment, the period for updating the parameters of the artificial neural network can be configured to correspond to information associated with the moving speed of the electronic device.

[0130] An embodiment provides a method for estimating the characteristics of a wireless communication channel by an electronic device. The method may include: receiving signals from an external electronic device through at least one of a plurality of antennas of the electronic device. The method may include: identifying whether information associated with the characteristics of the wireless communication channel identified from the received signals satisfies conditions for identifying the reliability of training data. The method may include: performing online training of an artificial neural network for estimating the characteristics of a wireless communication channel corresponding to a cell, using the information associated with the characteristics of the wireless communication channel, based on the satisfaction of the conditions. The method may include: estimating the characteristics of the wireless communication channel based on information output from the online-trained artificial neural network by inputting signals received through at least a portion of the plurality of antennas.

[0131] In an embodiment, the method may further include: obtaining the instantaneous PDP and instantaneous Doppler extension based on the information output by the untrained artificial neural network by inputting the received signal into the artificial neural network that has not undergone online training.

[0132] In an embodiment, the method may further include: determining the information associated with the estimation result of the wireless communication channel as the true value of the training data for training the artificial neural network based on the condition that the information associated with the estimation result of the wireless communication channel satisfies the identification.

[0133] In an embodiment, the method may further include: identifying whether a condition for updating the parameters of the artificial neural network is met. The method may further include: fine-tuning the artificial neural network by updating the parameters based on the identified condition being met.

[0134] In an embodiment, the method may further include: performing fine-tuning of the artificial neural network based on the period used to update the parameters of the artificial neural network.

[0135] An embodiment may provide a storage medium storing at least one instruction readable by a computer. When executed by at least one processor of an electronic device, the at least one instruction can cause the electronic device to perform at least one operation. The at least one operation may include: receiving a signal from an external electronic device through at least one of a plurality of antennas of the electronic device. The at least one operation may include: identifying whether information associated with characteristics of a wireless communication channel identified from the received signal satisfies conditions for identifying the reliability of training data. The at least one operation may include: performing online training of an artificial neural network for estimating the characteristics of a wireless communication channel corresponding to a cell, using the information associated with the characteristics of the wireless communication channel, based on the satisfied condition. The at least one operation may include: estimating the characteristics of the wireless communication channel based on information output from the online-trained artificial neural network by inputting signals received through at least a portion of the plurality of antennas.

Claims

1. An electronic device, comprising: Multiple antennas; Memory; as well as At least one processor, Wherein, the at least one processor is configured to: Signals are received from an external electronic device via at least one of the plurality of antennas; Whether the information associated with the characteristics of the wireless communication channel identified from the received signal meets the conditions for the reliability of the training data used for identification; Based on the information identified as being associated with the characteristics of the wireless communication channel and satisfying the condition, online training of an artificial neural network for estimating the characteristics of the wireless communication channel corresponding to a cell is performed using the information associated with the characteristics of the wireless communication channel; and The characteristics of the wireless communication channel are estimated based on information from the output of the online-trained artificial neural network by inputting signals received through at least a portion of the plurality of antennas.

2. The electronic device according to claim 1, wherein, The information associated with the characteristics of the wireless communication channel includes instantaneous power delay distribution (PDP) and instantaneous Doppler spread, and The at least one processor is further configured to: obtain the instantaneous PDP and the instantaneous Doppler extension by inputting the received signal into an artificial neural network that has not performed the online training, based on information from the output of the artificial neural network that has never performed the online training.

3. The electronic device according to claim 1, wherein, The conditions used to identify the reliability of the training data include the following: the cyclic redundancy check (CRC) result of the packets decoded from the received signal corresponds to a value indicating that there are no errors in the decoded packets.

4. The electronic device according to claim 1, wherein, The conditions used to identify the reliability of the training data include the following: the CRC result of the packet decoded from the received signal corresponds to a value indicating that there is an error in the decoded packet, and the signal-to-noise ratio (SNR) of the received signal exceeds a threshold.

5. The electronic device according to claim 1, wherein, The conditions used to identify the reliability of the training data include the following: the CRC result of the packet decoded from the received signal corresponds to a value indicating that there are no errors in the decoded packet, and the SNR of the received signal exceeds a threshold.

6. The electronic device according to claim 1, wherein, The at least one processor is further configured to: determine the information associated with the estimation result of the wireless communication channel as the true value of training data for training the artificial neural network based on the condition that the information associated with the estimation result of the wireless communication channel satisfies the condition.

7. The electronic device according to claim 1, wherein, The at least one processor is further configured to: To determine whether the conditions for updating the parameters of the artificial neural network are met; and Based on the recognition that the condition is met, the artificial neural network is fine-tuned by updating the parameters.

8. The electronic device according to claim 7, wherein, The conditions used to update the parameters of the artificial neural network include the following: the quality of the received signal is below a threshold.

9. The electronic device according to claim 1, wherein, The at least one processor is further configured to perform fine-tuning of the artificial neural network based on the period used to update the parameters of the artificial neural network.

10. The electronic device according to claim 9, wherein, The period used to update the parameters of the artificial neural network is configured to correspond to information associated with the moving speed of the electronic device.

11. A method for estimating the characteristics of a wireless communication channel by an electronic device, the method comprising: Signals are received from an external electronic device through at least one of the multiple antennas of the electronic device; Whether the information associated with the characteristics of the wireless communication channel identified from the received signal meets the conditions for the reliability of the training data used for identification; Based on the information identified as being associated with the characteristics of the wireless communication channel, the conditions are satisfied, and the information associated with the characteristics of the wireless communication channel is used to perform online training of an artificial neural network for estimating the characteristics of the wireless communication channel corresponding to the cell. as well as The characteristics of the wireless communication channel are estimated based on information from the output of the online-trained artificial neural network by inputting signals received through at least a portion of the plurality of antennas.

12. The method according to claim 11, wherein, The information associated with the characteristics of the wireless communication channel includes instantaneous PDP and instantaneous Doppler spread, and The method further includes: by inputting the received signal into an artificial neural network that has not undergone online training, and obtaining the instantaneous PDP and the instantaneous Doppler extension based on the information output by the artificial neural network that has never undergone online training.

13. The method according to claim 11, wherein, The conditions used to identify the reliability of the training data include the following: the CRC result of the packet decoded from the received signal corresponds to a value indicating that there are no errors in the decoded packet.

14. The method according to claim 11, wherein, The conditions used to identify the reliability of the training data include the following: the CRC result of the packet decoded from the received signal corresponds to a value indicating that there is an error in the decoded packet, and the SNR of the received signal exceeds a threshold.

15. A storage medium storing at least one instruction readable by a computer, wherein, When the at least one instruction is executed by at least one processor of the electronic device, it causes the electronic device to perform at least one operation, and Wherein, the at least one operation includes: Signals are received from an external electronic device through at least one of the multiple antennas of the electronic device; Whether the information associated with the characteristics of the wireless communication channel identified from the received signal meets the conditions for the reliability of the training data used for identification; Based on the information identified as being associated with the characteristics of the wireless communication channel and satisfying the condition, online training of an artificial neural network for estimating the characteristics of the wireless communication channel corresponding to a cell is performed using the information associated with the characteristics of the wireless communication channel; and The characteristics of the wireless communication channel are estimated based on information from the output of the online-trained artificial neural network by inputting signals received through at least a portion of the plurality of antennas.

Images