Method and device for transmitting data for training artificial intelligence model in wireless communication system

Abstract

A method for operating a first electronic device according to embodiments of the present disclosure comprises the steps of: transmitting, to a second electronic device, information for generating label data including at least one bit value; instructing the second electronic device to transmit the label data; receiving a signal including the label data from the second electronic device; processing the received signal to obtain input data for training an artificial intelligence model; and training the artificial intelligence model on the basis of the input data and the label data.

Description

Data transmission method and device for training an artificial intelligence model in a wireless communication system

[0001] The present disclosure relates to a method and apparatus for transmitting data to train an artificial intelligence model in a wireless communication system. Specifically, it relates to a method and apparatus that can utilize a received signal as training data for an artificial intelligence model regardless of the performance of a receiver in a wireless communication system.

[0002] Looking back at the evolution of wireless communication through successive generations, technologies have been developed primarily for human-oriented services, such as voice, multimedia, and data. Following the commercialization of 5G (5th Generation) communication systems, connected devices, which have been increasing explosively, are expected to be connected to communication networks. Examples of networked objects include vehicles, robots, drones, home appliances, displays, smart sensors installed in various infrastructures, construction machinery, and factory equipment. Mobile devices are expected to evolve into various form factors, such as augmented reality glasses, virtual reality headsets, and holographic devices. In the 6G (6th Generation) era, efforts are underway to develop improved 6G communication systems to connect hundreds of billions of devices and objects to provide diverse services. For this reason, 6G communication systems are being referred to as "beyond 5G" systems.

[0003] In the 6G communication system predicted to be realized around 2030, the maximum transmission speed is tera (i.e., 1,000 gigabit) bps (bit per second), and the wireless latency is 100 microseconds (μsec). In other words, compared to the 5G communication system, the transmission speed in the 6G communication system is 50 times faster, and the wireless latency is reduced to one-tenth.

[0004] To achieve such high data transmission speeds and ultra-low latency, 6G communication systems are being considered for implementation in the terahertz (THz) band (e.g., the 95 gigahertz (GHz) to 3 terahertz (3THz) band). Due to more severe path loss and atmospheric absorption phenomena compared to the millimeter wave (mmWave) band introduced in 5G, the importance of technologies capable of guaranteeing signal reach, or coverage, is expected to increase in the terahertz band. As key technologies to ensure coverage, new waveforms, beamforming, and multi-antenna transmission technologies such as massive Multiple-Input and Multiple-Output (MIMO), Full Dimensional MIMO (FD-MIMO), array antennas, and large-scale antennas, which are superior in terms of coverage compared to RF (Radio Frequency) devices, antennas, and OFDM (Orthogonal Frequency Division Multiplexing), must be developed. In addition, new technologies such as metamaterial-based lenses and antennas, high-dimensional spatial multiplexing technology using Orbital Angular Momentum (OAM), and Reconfigurable Intelligent Surface (RIS) are being discussed to improve the coverage of terahertz band signals.

[0005] In addition, to improve frequency efficiency and system network, development is underway in 6G communication systems for full duplex technology, in which uplink and downlink simultaneously utilize the same frequency resources at the same time; network technology that integrates satellites and HAPS (High-Altitude Platform Stations); network structure innovation technology that supports mobile base stations and enables network operation optimization and automation; dynamic spectrum sharing technology through collision avoidance based on spectrum usage prediction; AI-based communication technology that utilizes AI (Artificial Intelligence) from the design stage and internalizes end-to-end AI support functions to realize system optimization; and next-generation distributed computing technology that realizes services of complexity exceeding the limits of terminal computing capabilities by utilizing ultra-high performance communication and computing resources (Mobile Edge Computing (MEC), cloud, etc.). In addition, attempts are continuing to further strengthen connectivity between devices, further optimize networks, promote the softwareization of network entities, and increase the openness of wireless communication through the design of new protocols to be used in 6G communication systems, the implementation of hardware-based security environments, the development of mechanisms for the safe utilization of data, and the development of technologies regarding privacy maintenance methods.

[0006] Due to the research and development of such 6G communication systems, it is expected that a new dimension of hyper-connected experience will become possible through the hyper-connectivity of 6G communication systems, which encompasses not only connections between objects but also connections between people and objects. Specifically, it is projected that 6G communication systems will enable the provision of services such as truly immersive eXtended Reality (XR), high-fidelity mobile holograms, and digital replicas. Furthermore, services such as remote surgery, industrial automation, and emergency response, which are provided through 6G communication systems with enhanced security and reliability, will be applied in various fields including industry, healthcare, automotive, and home appliances.

[0007] One objective of the present disclosure may be to provide a method and apparatus for transmitting data for training an artificial intelligence (AI) model or a machine learning (ML) model (AI / ML model) in communication systems such as 5G, 5G-Advanced, and 6G.

[0008] A method performed by a first electronic device in a wireless communication system according to embodiments of the present disclosure comprises: transmitting information to a second electronic device for generating label data including at least one bit value; instructing the second electronic device to transmit the label data; receiving a signal including the label data from the second electronic device; processing the received signal to obtain input data for training an artificial intelligence model; and training an artificial intelligence model based on the input data and the label data.

[0009] In various embodiments, the first electronic device may include a receiving end, and the second electronic device may include a transmitting end. The first electronic device and the second electronic device may each represent a terminal or a base station.

[0010] In various embodiments, information for generating label data may include at least one of first information (see FIG. 8) including a plurality of indexes corresponding to a plurality of label data, second information (see FIG. 9) including a plurality of indexes corresponding to a plurality of sequence generation methods, or third information regarding a randomly determined seed value.

[0011] In various embodiments, the third information may include information about a seed value, a formula for generating the seed value, and / or parameters applied to the formula for generating the seed value. Additionally, the formula for generating the seed value may be based on the number of symbols per slot, a slot number within the frame, an OFDM symbol number, a scrambling identifier, or a physical layer cell identifier.

[0012] In various embodiments, the step of directing the transmission of label data may be performed based on RRC (radio resource control) signaling, MAC (medium access control) CE (control element) signaling, or DCI (downlink control indicator) signaling.

[0013] In various embodiments, the step of the first electronic device directing the transmission of label data may include the step of transmitting at least one of an index directing label data, an index directing a sequence generation method for generating label data, or a seed value for generating label data.

[0014] In various embodiments, the step of a first electronic device directing the transmission of label data further includes the step of allocating time resources and frequency resources for the transmission of label data, and the step of transmitting information about the time resources and frequency resources to a second electronic device, and the step of receiving a signal including label data can receive the signal based on the time resources and frequency resources.

[0015] In various embodiments, the step of a first electronic device receiving a signal containing label data may include identifying the location of the reference signal from the received signal based on the correlation between the reference signal and the received signal, and identifying the location of a signal corresponding to the label data from the received signal based on the location of the reference signal.

[0016] In various embodiments, the step of the first electronic device identifying the location of a reference signal can identify the location of the reference signal from the received signal based on the correlation between the signal corresponding to the reference signal and the label data and the received signal.

[0017] The method and apparatus according to the embodiments of the present disclosure can provide an operation for effectively transmitting data for training an AI model in a wireless communication system.

[0018] Specifically, the embodiments of the present disclosure allow the receiver to know the processing result of the received signal and acquire a data set for training an artificial intelligence model, even if the receiver cannot directly process the received signal. Accordingly, the problem of not being able to use the data for training the artificial intelligence model due to the receiver's inability to process the received signal is resolved, and the problem of the performance of the trained artificial intelligence model being dependent on the receiver's performance is resolved.

[0019] The effects obtainable from the present disclosure are not limited to those mentioned above, and other unmentioned effects will be clearly understood by those skilled in the art to which the present disclosure belongs from the description below.

[0020] The features and advantages of the embodiments of the present disclosure will become more apparent from the following description together with the accompanying drawings.

[0021] FIG. 1 illustrates a wireless communication system according to embodiments of the present disclosure.

[0022] FIG. 2 is a drawing for explaining the structure of a terminal according to embodiments of the present disclosure.

[0023] FIG. 3 is a drawing for explaining the structure of a network entity (or base station) according to embodiments of the present disclosure.

[0024] FIG. 4 illustrates a data processing process according to embodiments of the present disclosure.

[0025] FIG. 5 shows an artificial intelligence model for reception applied to the data processing process of a receiving end according to embodiments of the present disclosure.

[0026] FIG. 6 shows an artificial intelligence model for reception applied to the data processing process of a receiving end according to embodiments of the present disclosure.

[0027] FIG. 7 illustrates a data transmission method for learning an artificial intelligence model according to embodiments of the present disclosure.

[0028] FIG. 8 shows information for generating label data for training an artificial intelligence model according to embodiments of the present disclosure.

[0029] FIG. 9 illustrates sequence generation methods related to label data for training an artificial intelligence model according to embodiments of the present disclosure.

[0030] FIG. 10 illustrates a synchronization process for collecting data for training an artificial intelligence model according to embodiments of the present disclosure.

[0031] FIG. 11 illustrates a data transmission method for learning an artificial intelligence model according to embodiments of the present disclosure.

[0032] FIG. 12 illustrates a data transmission method for learning an artificial intelligence model according to embodiments of the present disclosure.

[0033] FIG. 13 shows input data and output data for training an artificial intelligence model according to embodiments of the present disclosure.

[0034] FIG. 14 shows input data and output data for training an artificial intelligence model according to embodiments of the present disclosure.

[0035] Embodiments of the present disclosure may solve the problems and / or disadvantages described above and provide the advantages described below. One aspect of the present disclosure may provide a network entity (or node) and a method of communication thereof in a wireless communication system.

[0036] The terms used in this disclosure are used merely to describe specific embodiments and are not intended to limit the scope of other embodiments. A singular expression may include a plural expression unless the context clearly indicates otherwise. Terms used herein, including technical or scientific terms, may have the same meaning as generally understood by those skilled in the art described in this disclosure. Terms used in this disclosure that are defined in a general dictionary may be interpreted as having the same or similar meaning as they have in the context of the relevant technology, and are not to be interpreted in an ideal or overly formal sense unless explicitly defined in this disclosure. In some cases, even terms defined in this disclosure are not to be interpreted to exclude the embodiments of this disclosure.

[0037] The various embodiments of the present disclosure described below illustrate a hardware-based approach. However, since the various embodiments of the present disclosure include techniques using both hardware and software, the various embodiments of the present disclosure do not exclude a software-based approach.

[0038] Additionally, various embodiments of the present disclosure describe various embodiments using terms used in some communication standards (e.g., 3GPP (3rd generation partnership project)), but this is merely for illustrative purposes. Various embodiments of the present disclosure can be easily modified and applied to other communication systems.

[0039] Various embodiments of the present disclosure are described below.

[0040] FIG. 1 illustrates a wireless communication system according to embodiments of the present disclosure.

[0041] FIG. 1 illustrates a base station (110), a first terminal (120), and / or a second terminal (130) as part of nodes utilizing a wireless channel in a wireless communication system. FIG. 1 illustrates only one base station, but this is merely an example. The wireless communication system of FIG. 1 may include other base stations identical or similar to the base station (110).

[0042] A base station (110) is a network infrastructure that provides wireless access to terminals (120, 130). The base station (110) has coverage defined as a certain geographical area based on the distance at which it can transmit signals. In addition to being a base station, the base station (110) may be referred to as an 'access point (AP)', 'evolved Node B (eNB)', 'next generation node B (gNB)', '5G node (5th generation node)', 'wireless point', 'transmission / reception point (TRP)', or other terms having an equivalent technical meaning.

[0043] Each of the first terminal (120) and the second terminal (130) is a device used by a user and can perform communication with the base station (110) via a wireless channel. At least one of the first terminal (120) or the second terminal (130) can be operated without user involvement. For example, at least one of the first terminal (120) or the second terminal (130) may be a device that performs machine type communication (MTC) and may not be carried by the user. Each of the first terminal (120) and the second terminal (130) may be referred to as 'user equipment (UE)', 'mobile station', 'subscriber station', 'customer premises equipment (CPE)', 'remote terminal', 'wireless terminal', 'electronic device', or 'user device' or other terms having an equivalent technical meaning.

[0044] The base station (110), the first terminal (120), and the second terminal (130) can transmit and / or receive wireless signals in a millimeter wave (mmWave) band (e.g., 28 GHz, 30 GHz, 38 GHz, 60 GHz). At this time, to improve channel gain, the base station (110), the first terminal (120), and / or the second terminal (130) can perform beamforming.

[0045] Beamforming may include transmitting beamforming and / or receiving beamforming. That is, the base station (110), the first terminal (120), and / or the second terminal (130) may give directivity to the transmitted signal or the received signal. To give directivity to the received signal, the base station (110) and / or the terminals (120, 130) may select serving beams (112, 113, 121, 131) through a beam search or beam management procedure. After the serving beams (112, 113, 121, 131) are selected, subsequent communication may be performed through a resource that is in a quasi-co-located (QCL) relationship with the resource that transmitted the serving beams (112, 113, 121, 131).

[0046] The base station (110), the first terminal (120), and the second terminal (130) of the present disclosure may each be a transmitting apparatus, a transmitting node, a receiving apparatus, and / or a receiving node. For example, the base station (110) may transmit a radio frequency (RF) signal to the first terminal (120). The base station (110) may receive an RF signal from the first terminal (120). As another example, the first terminal (120) may transmit an RF signal to the base station (110) or the second terminal (130). The first terminal (120) may receive an RF signal from the base station (110) or the second terminal (130).

[0047] FIG. 2 is a drawing for explaining the structure of a terminal according to embodiments.

[0048] Referring to FIG. 2, a terminal (200) according to embodiments may include a transceiver (transmitter / receiver) (210), a memory (220), and / or a processor (230). Although the present disclosure describes the terminal (200) as including a transceiver (210), a memory (220), and / or a processor (230), this is merely an example. For example, the terminal (200) may include additional components other than the transceiver (210), the memory (220), and the processor (230).

[0049] According to the embodiments, the transceiver (210), memory (220), and processor (230) may each be implemented or formed as separate chips. However, this is merely an example, and the transceiver (210), memory (220), and / or processor (230) may be implemented or formed as a single chip.

[0050] According to embodiments, the transceiver (210) may include at least one transmitter and / or at least one receiver. For example, the transceiver (210) may include an RF transmitter for amplifying and up-converting the frequency of a transmitted signal. The transceiver (210) may include an RF receiver for down-converting the frequency of a received signal and amplifying low-noise.

[0051] The configurations of the transceiver (210) described in this disclosure are merely examples and the configuration of the transceiver (210) is not limited to an RF transmitter and an RF receiver. For example, the transceiver (210) may further include a coupler to ensure isolation between the RF transmitter and the RF receiver.

[0052] According to the embodiments, the transceiver (210) can transmit or receive a signal to or from the processor (230). For example, the transceiver (210) can transmit or deliver an RF signal received through a wireless communication channel to or from the processor (230). The transceiver (210) can receive or receive an RF signal from or from the processor (230).

[0053] According to the embodiments, the transceiver (210) may be referred to as a UE transmitter or a UE receiver.

[0054] According to embodiments, the transceiver (210) may transmit a signal to a base station (e.g., base station (110) of FIG. 1) or a network entity (e.g., an access and mobility management function (AMF) entity) or receive a signal from a base station or a network entity. In embodiments, the transmitted or received signal may include control signals and data.

[0055] According to embodiments, the memory (220) may include or store programs and data necessary for the operations of the terminal (200). For example, the memory (220) may be a non-transitory memory, and a program stored in the non-transitory memory may be organically coupled with the hardware configuration of the terminal (200) (e.g., a processor (230) or a transceiver (210)). The memory (220) may store control information or data including signals obtained by the terminal (200). In embodiments, the memory (220) may include a read-only memory (ROM), a random access memory (RAM), a hard disk, a CD-ROM, a DVD, and / or a storage medium.

[0056] According to the embodiments, the processor (230) may include one processor or a plurality of processors. For example, the processor (230) may include a communication processor. For example, the processor (230) may include a communication processor and / or an application processor.

[0057] According to embodiments, the processor (230) can control a series of processes performed by the terminal (200). For example, the transceiver (210) can receive a data signal containing control information transmitted by a base station or network entity. The processor (230) can process the received control signal and data signal.

[0058] The term processor in the present disclosure may be replaced with various terms referring to a configuration that executes or performs operations of the terminal (200). For example, the processor may be replaced with a controller or a computing circuit.

[0059] The terminal (200) of the present disclosure may correspond to the first terminal (120) and / or the second terminal (130) of FIG. 1.

[0060] FIG. 3 is a diagram illustrating the structure of a network entity (or base station) according to embodiments.

[0061] Referring to FIG. 3, a network entity (300) according to embodiments may include a transceiver (transmitter / receiver) (310), a memory (320), and / or a processor (330). Although the present disclosure describes the network entity (300) as including a transceiver (310), a memory (320), and / or a processor (330), this is merely an example. For example, the network entity (300) may include additional components other than the transceiver (310), the memory (320), and the processor (330). The network entity (300) may represent network functions included in a base station or other core network.

[0062] According to the embodiments, the transceiver (310), memory (320), and processor (330) may each be implemented or formed as separate chips. However, this is merely an example, and the transceiver (310), memory (320), and / or processor (330) may be implemented or formed as a single chip.

[0063] According to embodiments, the transceiver (310) may include at least one transmitter and / or at least one receiver. For example, the transceiver (310) may include an RF transmitter for amplifying and up-converting the frequency of a transmitted signal. The transceiver (310) may include an RF receiver for down-converting the frequency of a received signal and amplifying low-noise.

[0064] The configurations of the transceiver (310) described in this disclosure are merely examples and are not limited to an RF transmitter and an RF receiver. For example, the transceiver (310) may further include a coupler to ensure isolation between the RF transmitter and the RF receiver.

[0065] According to the embodiments, the transceiver (310) can transmit or receive a signal to or from the processor (330). For example, the transceiver (310) can transmit or deliver an RF signal received through a wireless communication channel to or from the processor (330). The transceiver (310) can receive or receive an RF signal from the processor (230).

[0066] According to the embodiments, the transceiver (310) may be referred to as a network entity transmitter or a network entity receiver.

[0067] According to embodiments, the transceiver (310) can transmit a signal to the terminal (200) or receive a signal from the terminal (200). In embodiments, the transmitted or received signal may include control signals and data.

[0068] According to embodiments, the memory (320) may contain programs and data necessary for the operations of the network entity (300). For example, the memory (320) may be a non-transitory memory, and a program stored in the non-transitory memory may be organically coupled with the hardware configuration of the network entity (300) (e.g., a processor (330) or a transceiver (310)). The memory (320) may store control information or data including signals obtained by the network entity (300). In embodiments, the memory (320) may include read-only memory (ROM), random access memory (RAM), a hard disk, a CD-ROM, a DVD, and / or storage media.

[0069] According to the embodiments, the processor (330) may include one processor or a plurality of processors. For example, the processor (330) may include a communication processor. For example, the processor (330) may include a communication processor and / or an application processor.

[0070] According to embodiments, the processor (330) can control a series of processes performed by the network entity (300). For example, the transceiver (310) may receive a data signal containing control information transmitted by the network entity. The processor (330) may process the received control signal and data signal.

[0071] The term processor in the present disclosure may be replaced with various terms referring to a configuration that executes or performs operations of a network entity (300). For example, processor may be replaced with a controller or a computing unit.

[0072] The network entity (300) of the present disclosure may correspond to the base station (110) of FIG. 1.

[0073] The device described in FIGS. 2 and 3 may correspond to a device of a transmitting end or a receiving end. A terminal or network entity according to embodiments of the present disclosure may be a transmitting end when it is a transmitting end, and may be a receiving end when the terminal or network entity is a receiving end.

[0074] In the following, the transmitting end and the receiving end may respectively refer to the terminal or base station described in FIGS. 1 to 3. When describing a downlink signal, the base station will be the transmitting end and the terminal will be the receiving end, and when describing an uplink signal, the terminal will be the transmitting end and the base station will be the receiving end.

[0075] FIG. 4 illustrates the data processing process of a transmitting end and a receiving end according to embodiments of the present disclosure.

[0076] Referring to FIG. 4, a process is illustrated in which a transmitting end (410) processes bit information (e.g., encoding) and transmits a signal to a receiving end (420) through a channel, and the receiving end (420) processes the received signal (e.g., decoding) to restore the bit information.

[0077] In FIG. 4, the transmitting end (410) and the receiving end (420) may include a user terminal, a base station, or other network function entity.

[0078] In step 4102 of FIG. 4, the transmitting end (410) comprises a vector including a plurality of bits consisting of 0 or 1. Channel coding (4102) is performed on the channel-coded vector It can generate. Channel coding is a vector related to data error correction. An operation to insert additional bits may be indicated. Channel coding operations may be performed to increase the reliability of data transmission. Corresponding to step 4102, the receiving end (420) may perform channel decoding operations (4218) in reverse.

[0079] In step 4104, the transmitting end (410) is a vector Rate matching can be performed on the. Rate matching may represent an operation to adjust data into a form suitable for transmission, such as rearranging or deleting data after channel coding to adjust the transmission speed. Corresponding to step 4104, the receiving end (420) may perform a rate dematching operation (4216) in reverse.

[0080] In step 4106, the transmitting end (410) scrambles the vector generated after rate matching and vector It can generate. Scrambling may be an operation performed to mix data patterns to reduce interference or distortion. Scrambling may be used to minimize noise and interference and to increase security during data transmission. The transmitting end (410) may scramble data using a mathematical algorithm, and the receiving end (420) may perform a descrambling operation (4214) to restore the original data using the same algorithm.

[0081] In step 4108, the transmitting end (410) is a vector Modulating the data symbol It can generate. Modulation refers to one of the processes for converting digital data into an analog signal for transmission. Modulation refers to the process of converting an information signal into a transmittable signal (frequency, amplitude, phase, etc.). The modulated signal can be restored to the original signal through a demodulation process. Modulation methods may include amplitude shift keying (ASK), frequency shift keying (FSK), phase shift keying (PSK), quadrature amplitude modulation, etc. Corresponding to step 4108, the receiving end (420) can perform a demodulation operation (4212) in reverse.

[0082] In step 4110, the transmitting end (410) has data symbols Regarding this, a mapped data symbol g can be generated by mapping the layer and the RE (resource element). Layer and RE mapping can represent the operation of assigning data to the layer and the resource element. In MIMO (multiple input multiple output) technology, which transmits data simultaneously using multiple antennas, layer mapping can increase transmission efficiency by dividing the data and assigning it to various antennas. Additionally, RE mapping can optimize communication resources by effectively mapping data to time resources and frequency resources. Corresponding to step 4110, the receiving end (420) can perform a layer and RE demapping operation (4206) in reverse.

[0083] In various embodiments, step 4110 may be performed based on a reference signal such as DMRS (demodulation reference signal) or SRS (sounding reference signal).

[0084] In step 4112, the transmitting end (410) mapped data symbols Orthogonal Frequency Division Multiplexing (OFDM) modulation can be performed on the above. OFDM modulation is a frequency division multiplexing method that can represent a method of dividing data into multiple orthogonal subcarriers for transmission. The subcarriers are divided at regular intervals along the frequency axis, and since each is orthogonal to the others, efficient data transmission can be performed without interference. Corresponding to step 4112, the receiving end (420) can perform OFDM demodulation (4202) in reverse.

[0085] The description of the transmission process and the corresponding reception process of Fig. 4 is as follows.

[0086] In step 4202, the receiver (420) can perform OFDM demodulation (4202) to convert the received OFDM signal back into the original data.

[0087] In step 4204, the receiving unit (420) can restore the signal by performing STO (sampling time offset) and CFO (carrier frequency offset) correction (STO and CFO Correction) (4204) to correct the time error (STO) and frequency error (CFO) between the transmitting and receiving units.

[0088] In various embodiments, step 4204 may be performed based on a reference signal such as DMRS (demodulation reference signal) or SRS (sounding reference signal).

[0089] In step 4206, the receiving end (420) can perform Layer and RE Demapping (4206) to separate the data again from the layer and resource element (RE) allocated at the time of transmission. The separated data It can be expressed as.

[0090] In step 4208, the receiving end (420) can estimate the channel state between the transmitting end and the receiving end through a channel estimator (4208). The receiving end (420) can perform accurate data restoration by correcting signal distortion based on the channel state estimated by the channel estimator.

[0091] In various embodiments, step 4208 may be performed based on a reference signal such as DMRS (demodulation reference signal) or SRS (sounding reference signal).

[0092] In step 4210, the receiving unit (420) can restore data by correcting channel distortion through a channel equalizer (4210). At this time, information regarding the channel state estimated from a channel estimator may be utilized. The data restored with channel distortion corrected It can be expressed as a data symbol processed at the transmitting end. It can respond to.

[0093] In step 4212, the receiving unit (420) may perform soft demodulation (4212). Soft demodulation is one of the processes for converting a received signal into digital bits and may represent a method of estimating bit values ​​using likelihood information of the received signal. For example, soft demodulation can improve decoding performance through a value representing the probability that each bit of the data is 0 or 1. The data on which soft demodulation has been performed It can be expressed as, the vector after scrambling processed at the transmitting end. It can respond to.

[0094] In step 4214, the receiving unit (420) is Descrambling (4214) can be performed on it. The receiving end (420) can restore the original data pattern by releasing the scrambling applied during the transmission process.

[0095] In step 4216, the receiving unit (420) can perform rate dematching (4216) on the descrambled data. The receiving unit (420) can restore the data to its original length by performing the rate matching process applied during the transmission process in reverse. At this time, the restored data It can be expressed as, and is a vector processed at the transmitting end. It can respond to.

[0096] In step 4218, the receiving unit (420) can perform channel decoding (4218). The receiving unit (420) can restore the original data bits with error correction by performing the channel coding process applied during the transmission process in reverse. At this time, the restored data It can be expressed as.

[0097] Represents the bit information to be transmitted during the transmission process and restored at the receiving end Depending on the channel state, it may be completely identical or differ in some aspects.

[0098] In various embodiments, is TX uncoded data bits, can be referred to as RX uncoded data bits. is TX coded data bits, can be referred to as RX coded data bits. is TX coded data bits after scrambling, This can be referred to as log likelihood ratio values ​​before descrambling. is TX modulated data symbols, can be referred to as RX modulated data symbols. TX mapped data symbols on resource grid, It can be referred to as RX mapped data symbols on resource grid.

[0099] FIG. 5 shows an artificial intelligence model for reception applied to the data processing process of a receiving end according to embodiments of the present disclosure.

[0100] Referring to FIG. 5, some steps of the process of handling received data at the receiver may be performed by an artificial intelligence (AI) model for reception. For example, the AI ​​model of FIG. 5 can replace soft demodulation, descrambling, rate dematching, and channel decoding operations. For training the AI ​​model, data with channel distortion corrected through a channel equalizer at the receiver It can be used as input to an artificial intelligence model, and the artificial intelligence model uses the restored bit information based on the aforementioned input information. It can output. The data for training such an artificial intelligence model is and original bit information It may include, and the artificial intelligence model Output by the input a is the original bit information Learning can be performed in a way that improves the accuracy of the output information by checking whether it is identical to or not. In this case, the original bit information is information used to compare whether the output value of the artificial intelligence model is the correct answer or not, and may be referred to as label data, reference data, actual data, truth data, verification data, fact data, ground truth data, or other terms with similar meanings.

[0101] In various embodiments of the present disclosure, the steps of the reception procedure that can be replaced by an artificial intelligence model for reception may vary, and accordingly, the types of input data and label data in the data for training said artificial intelligence model may vary. Therefore, even if not all cases are described in the present disclosure, it can be understood that various cases of artificial intelligence models and the types of data required to train said artificial intelligence model are described.

[0102] In various embodiments, the label data may include not only bits but also symbols or signals that can be generated from the bits.

[0103] FIG. 6 shows an artificial intelligence model for reception applied to the data processing process of a receiving end according to embodiments of the present disclosure.

[0104] Referring to FIG. 6, some steps of the process of handling received data at the receiver may be performed by an artificial intelligence model for reception. For example, the artificial intelligence model of FIG. 6 can replace channel equalizer and soft demodulation operations. For training the artificial intelligence model, data generated through layer and RE demodulation steps at the receiver It can be used as input to an artificial intelligence model, and the artificial intelligence model uses the restored data based on the aforementioned input information. It can output. Therefore, in FIG. 6, the AI ​​model for reception can replace the channel equalization step and the soft demodulation step, and the data output by the AI ​​model It can correspond to data after channel equalization and soft demodulation have been performed in the existing receiver.

[0105] The data for training the artificial intelligence model illustrated in Fig. 6 is and scrambled data during the transmission process It may include, and the artificial intelligence model Output by the input go Learning can be performed in a way that improves the accuracy of the output information by checking whether it is identical or not. At this time, Information used to compare whether the output value of an artificial intelligence model is correct or not may be referred to as label data, reference data, actual data, truth data, verification data, factual data, ground truth data, or other terms with similar meanings.

[0106] Meanwhile, although data required for training an artificial intelligence model requires label data, the receiver receiving the signal can only produce restored data during the signal restoration process and cannot accurately know the data actually transmitted by the transmitter (e.g., label data). Therefore, when training an artificial intelligence model using the signal received at the receiver, the model learns by treating the restored data, which may differ from the actual data, as the correct answer; consequently, even if the accuracy of the training results is high, the model's performance may be dependent on the existing data processing performance of the receiver. Furthermore, depending on the performance of the receiver, there may be a problem where the efficiency of training is reduced because the data of the corresponding signal cannot be used for training due to the lack of label data for data that the receiver cannot restore.

[0107] The method and apparatus according to the embodiments of the present disclosure have the effect of enabling an artificial intelligence model to be trained effectively and with high accuracy through label data at the receiving end.

[0108] FIG. 7 illustrates a data transmission method for learning an artificial intelligence model according to embodiments of the present disclosure.

[0109] Referring to FIG. 7, a transmission and reception operation between a terminal (710) and a base station (720) is illustrated. In FIG. 7, the terminal (710) may be a transmitting unit that transmits label data, and the base station (720) may be a receiving unit that receives a signal containing label data.

[0110] In step 732, the base station (720) may transmit information related to the generation of label data for the training of an artificial intelligence model to the terminal (710). Based on the information related to the generation of label data, the terminal (710) may generate label data. At this time, the label data may represent bit information including at least one bit. That is, the label data may represent a vector including bits composed of 0 or 1.

[0111] Information related to the generation of label data may include various information that enables the terminal (710) to generate label data by the instruction of the base station (720). For example, information related to the generation of label data may include at least one of table information including multiple bit information and corresponding indices (see FIG. 8), table information including indices corresponding to sequence generation methods for generating label data (see FIG. 9), or information regarding a seed value for generating label data.

[0112] In step 734, the base station (720) may transmit a message to instruct the terminal (710) to transmit label data. That is, the base station (720) may instruct the terminal (710) to transmit label data. At this time, the instruction may be performed based on RRC (radio resource control) signaling, MAC (medium access control) CE (control element) signaling, or DCI (downlink control indicator) signaling.

[0113] In various embodiments, persistent or semi-persistent instructions may be transmitted from a base station to a terminal based on RRC signaling. Additionally, dynamic instructions may be transmitted from a base station to a terminal in a manner that activates or deactivates the transmission of label data based on MAC CE signaling or DCI signaling.

[0114] In various embodiments, the operation of instructing to transmit label data may represent an operation of instructing a separate test terminal or an operation of instructing a terminal for actual use.

[0115] In various embodiments, the action of instructing to transmit label data may represent an action of instructing to transmit label data of a specific size at a specific time. For example, it may represent instructing to transmit label data for 10 slots at the top of every hour, or instructing to transmit label data during the evening hours when data traffic is low.

[0116] In various embodiments, the operation of instructing to transmit label data may be selectively instructed to at least some of the multiple terminals, taking into account the terminals' mobility, location information, battery or data requirements, etc.

[0117] In various embodiments, instructions for transmitting label data may include at least one of an index corresponding to specific bit information, an index corresponding to a specific sequence generation method, or information regarding a seed value for generating a specific sequence, based on information related to the generation of label data. Additionally, instructions for transmitting label data may include parameters for generating label data.

[0118] In various embodiments, information regarding a seed value for generating a specific sequence may include the seed value, information instructing the use of a previously transmitted seed value (e.g., when a seed value is transmitted in a 732 operation), information instructing the derivation and use of a seed value using previously transmitted information (e.g., when transmitted in a 732 operation) (e.g., a mathematical formula or parameter for generating a seed value).

[0119] In step 736, the terminal (710) may transmit a signal containing label data to the base station (720). The terminal (710) may identify the label data based on the information received from the base station (720) in step 732 or step 734. The terminal (710) may identify the label data based on the index received in step 734, or identify a sequence capable of generating the label data, and may generate the label data by generating a seed value based on the information received in step 734. Then, the terminal (710) may process the generated or identified label data according to the data processing method described in FIG. 4 and transmit it to the base station (720).

[0120] Likewise, in step 736, the base station (720) may receive a signal containing label data from the terminal (710). At this time, the base station (720) may identify the location of the reference signal received from the terminal based on the correlation of the reference signal to identify the signal corresponding to the label data, and may identify the location of the signal corresponding to the label data from the identified location of the reference signal.

[0121] The base station (720) compares the pattern of the received signal with the pattern of the reference signal, and if the correlation is greater than a specific value, identifies the signal pattern as the reference signal, and can identify the location of the signal corresponding to the label data based on the identified reference signal. Such an operation may be referred to as synchronization.

[0122] In various embodiments, the base station (720) can identify the location of a received reference signal based on the correlation between a reference signal pattern and a signal pattern corresponding to label data and a received signal pattern, and can identify the location of a signal corresponding to label data from the identified location of the reference signal. That is, the base station (720) can use the reference signal pattern and the signal pattern corresponding to label data together when determining the correlation. When confirming the correlation with the received signal by using both the reference signal pattern and the signal pattern corresponding to label data, there is an effect of being able to identify the location of the reference signal more accurately even if there is signal distortion and noise.

[0123] In various embodiments, when synchronization is ensured between the terminal (710) and the base station (720) based on TA (timing advance) information, the operation of identifying the location of the received reference signal based on correlation may be omitted.

[0124] In various embodiments, the terminal (710) receives information about resources (time resources and frequency resources) allocated (or scheduled) for transmitting label data from the base station (720), and can transmit a signal containing label data to the base station (720) based on the allocated resources.

[0125] In various embodiments, label data may be scheduled simultaneously with data containing other information (e.g., allocated to the same time resource) or scheduled to different time resources. For example, in the case of a terminal requiring high-volume uplink transmission, data containing actual information may be allocated to the same time resource as the label data for data transmission.

[0126] In step 738, the base station (720) can process a signal received from the terminal (710) to obtain data for training an artificial intelligence model. For example, based on the data processing method of the receiving end described in FIG. 4, various data corresponding to input information for training an artificial intelligence model (e.g., , You can obtain (etc.).

[0127] In step 740, the base station (720) can train an artificial intelligence model based on the artificial intelligence model training data collected in the aforementioned step 738. The artificial intelligence model training data can be configured in various ways according to the received data processing procedure to be replaced by the artificial intelligence model. For example, in the case of an artificial intelligence model to replace soft demodulation (4212), descrambling (4214), rate dematching (4216), and channel decoding (4218) operations as shown in FIG. 4, the artificial intelligence model training data is data generated from a channel equalizer as input data. Data that may include, and includes a bit value of 0 or 1 as label data It may include. Additionally, for example, in the case of an artificial intelligence model to replace the operation of a channel equalizer (4210) and soft demodulation (4212) as in FIG. 6, the data for training the artificial intelligence model may include layer and RE-demapped data as input data. It may include, and data as label data It may include.

[0128] In various embodiments, the label data is bit data composed of bit values ​​of 0 or 1. Data generated by processing from (e.g., It may include (etc.). The base station (720) can configure the necessary label data according to the type of artificial intelligence model.

[0129] In various embodiments, the input data and label data for training the artificial intelligence model may be configured differently depending on the type of artificial intelligence model.

[0130] FIG. 8 shows information for generating label data for training an artificial intelligence model according to embodiments of the present disclosure.

[0131] Referring to FIG. 8, a table including a plurality of bit data and corresponding indices is illustrated as an example. Here, the bit data may include a vector containing bit values ​​of 0 or 1, and the bit data is the data described in FIG. 4. It can correspond to. Also, bit data can correspond to label data. Bit data may also be referred to as a bit sequence.

[0132] In various embodiments, the information exemplified in FIG. 8 may include bit data and corresponding indices configured differently according to the number of resource blocks (RB), the number of slots, the modulation coding scheme (MCS), and / or the type of OFDM (e.g., Cyclic prefix-OFDM or Dft-s-OFDM, etc.). That is, the information for generating label data for training an artificial intelligence model may include a plurality of tables according to the aforementioned conditions. The bit data may be composed of a sequence capable of generating a signal representative of the overall signal by considering statistical characteristics of the signal, such as size distribution.

[0133] In various embodiments, the sequence length of bit data may be determined based on the number of RBs, the number of slots, MCS, etc. For example, the sequence length (or the number of bits) may be calculated by the following [Equation 1].

[0134] [Mathematical Formula 1]

[0135] Number of bits = (Number of REs) x (Modulation order) / (Code rate)

[0136] In various embodiments, the table information exemplified in FIG. 8 may be transmitted from a base station to a terminal by operation 732 of FIG. 7. And, in operation 734 of FIG. 7, the base station may transmit an index to the terminal to indicate bit data. Accordingly, the terminal can identify the bit data (or label data, bit sequence, bit information, etc.) indicated by the index received in operation 734 based on the table information received in operation 732.

[0137] And, in operation 736 of FIG. 7, the terminal can process the corresponding bit data based on the procedure performed by the transmitting unit (410) of FIG. 4 and transmit a signal containing the corresponding bit data to the base station. Since the base station is the entity that itself directed the label data, it already knows the label data. In addition, since the base station can obtain input data for training the artificial intelligence model through the signal received from the terminal, it can train the artificial intelligence model based on the label data and the input data.

[0138] In various embodiments, the base station may transmit a plurality of indices (e.g., a set of indices) to the terminal. The terminal may sequentially process label data corresponding to each of the plurality of indices and transmit a signal containing the label data to the base station.

[0139] In various embodiments, multiple bit sequences having a specific number of bits may each be indicated by an index. Accordingly, the terminal can identify which bit sequence is instructed to be transmitted based on the index included in the instruction received from the base station.

[0140] The table information exemplified in FIG. 8 may be referred to as a lookup table, and may also be referred to in various forms of information such as a list or an array. Any information containing multiple bit sequences and an identifier or indicator corresponding to each bit sequence can be understood as described by the example of FIG. 8.

[0141] FIG. 9 illustrates sequence generation methods related to label data for training an artificial intelligence model according to embodiments of the present disclosure.

[0142] Referring to FIG. 9, a table including a plurality of sequence generation methods and corresponding indices is illustrated as an example. Here, the sequence generation methods may include a pseudorandom binary sequence (e.g., gold sequence), Hadamard sequence, Fibonacci sequence, etc., and other known sequence generation methods may be additionally included.

[0143] Information for generating label data for training an artificial intelligence model may include a table containing multiple sequence generation methods and corresponding indices.

[0144] In various embodiments, the table information exemplified in FIG. 9 may be transmitted from the base station to the terminal by operation 732 of FIG. 7. And, in operation 734 of FIG. 7, the base station may transmit an index and / or a seed value for sequence generation to instruct the terminal on a sequence generation method. Accordingly, the terminal can identify the sequence generation method indicated by the index received in operation 734 based on the table information received in operation 732, and can generate label data based on the identified sequence generation method and the received seed value.

[0145] In addition, in operation 736 of FIG. 7, the terminal can process the generated label data based on the procedure performed by the transmitting unit (410) of FIG. 4 and transmit a signal containing the label data to the base station. Since the base station can generate the label data itself using the same sequence generation method and seed value as the terminal, it can know the label data. Furthermore, since the base station can obtain input data for training an artificial intelligence model through the signal received from the terminal, it can train the artificial intelligence model based on the label data and the input data.

[0146] In various embodiments, the base station may transmit a plurality of indices to the terminal. The terminal may generate label data by combining a sequence generation method corresponding to each of the plurality of indices.

[0147] In various embodiments, the seed value applied to the sequence generation method for generating label data may be specified as a sufficiently large number to ensure data diversity.

[0148] In various embodiments, the seed value used to generate label data may be set to a value that is commonly identifiable by the base station and the terminal. For example, a C_init value that is commonly identifiable by the base station and the terminal and has the characteristic of being randomly determined may be used as the seed value. The C_init value is 3GPP(3 rd It is a value disclosed in the TS (technical specification) 38.211 standard specification document of the generation partnership project, and can be calculated by [Equation 2].

[0149] [Mathematical Formula 2]

[0150]

[0151] : Indicates the number of symbols per slot.

[0152] : Can indicate the number of slots within the frame.

[0153] : Indicates the OFDM symbol number within the slot.

[0154] : Can represent scrambling identifiers for antenna ports.

[0155] : (Physical layer cell identifier) ​​or DMRS-scrambling-ID, pi2BPSK-scramblingID0 or pi2BPSK-scramblingID1.

[0156] In various embodiments, the seed value used to generate label data may be generated through a calculation utilizing parameters having random characteristics. In this case, the parameters having random characteristics may include parameters that can be commonly identified by the base station and the terminal. For example, the parameters having random characteristics may include an OFDM symbol number within a slot, a slot number within a frame, or a PRB (physical resource block) index.

[0157] In various embodiments, the seed value may be transmitted from the base station to the terminal in operation 732 or 734 of FIG. 7.

[0158] In various embodiments, information regarding a mathematical formula for calculating a seed value or a random parameter used to calculate a seed value in operation 732 of FIG. 7 may be transmitted from the base station to the terminal. In operation 734 of FIG. 7, the terminal receives instructions from the base station for transmitting label data, and the terminal calculates a seed value using a mathematical formula and / or parameter for calculating a seed value, and can generate label data according to an instructed sequence generation method using the calculated seed value.

[0159] In various embodiments, information about the sequence generation method and information about the seed value are transmitted to the transmitting end, and the transmitting end can generate label data using the indicated sequence generation method and seed value and transmit it to the receiving end.

[0160] In various embodiments, the base station may transmit a plurality of seed values ​​(e.g., a set of seed values) to the terminal. The terminal may sequentially process each calculated label data using the plurality of seed values ​​and transmit a signal containing the label data to the base station.

[0161] FIG. 10 illustrates a synchronization process for collecting data for training an artificial intelligence model according to embodiments of the present disclosure.

[0162] The receiver receives the signal sent by the transmitter and can locate the position of the signal containing label data from the received signal. The receiver can detect the position of the signal corresponding to the data for training an artificial intelligence model from the received signal, and a synchronization process can be performed to enable the receiver to detect the signal more accurately.

[0163] Referring to FIG. 10, the transmitter can transmit a signal to the receiver by OFDM modulating a reference signal (e.g., SRS (sounding reference signal), DM-RS (demodulation reference signal), PSS (primary synchronization signal), SSS (secondary synchronization signal), etc.) and a data signal (e.g., PUSCH (physical uplink shared channel), PDSCH (physical downlink shared channel), etc.) allocated on a grid of resources (time resources and frequency resources). FIG. 10 exemplarily illustrates the pattern of a signal transmitted to the receiver with resources allocated to the reference signal and the data signal.

[0164] The receiver receives a signal and can identify the location of a reference signal within the pattern of the received signal. For example, by comparing the pattern of the received signal with the pattern of the reference signal, the receiver can derive a value indicating the correlation between the two patterns; if this value is higher than a specific threshold, it can identify that a portion of the signal exhibiting high correlation is the reference signal. Once the receiver identifies the location of the reference signal from the received signal, it can easily identify the location of the data signal based on the location of the reference signal. Consequently, the receiver can process the identified data signal to obtain data for training an artificial intelligence model.

[0165] In various embodiments, the receiver can derive a value indicating the correlation between the two patterns by comparing the pattern of the received signal with the pattern of the signal containing the reference signal and label data, and if the value is higher than a specific threshold, it can identify that a portion of the signal showing high correlation is the signal containing the reference signal and label data. That is, by comparing the pattern of the signal containing not only the reference signal but also the pattern of the signal containing label data with the pattern of the received signal, the receiver can more accurately identify the location of the signal containing the reference signal and label data even in the presence of signal distortion and noise. This is because it is easier to detect signals with high correlation as the length of the signal increases.

[0166] In various embodiments, when a timing advance (TA) value is exchanged between a transmitting end and a receiving end to ensure synchronization, the operation of the receiving end comparing a pattern with a received signal with a reference signal or data signal may be omitted. The receiving end can identify the location of a signal containing label data based on the TA value.

[0167] FIG. 11 illustrates a data transmission method for learning an artificial intelligence model according to embodiments of the present disclosure.

[0168] Referring to FIG. 11, a transmission and reception operation between a terminal (1110) and a base station (1120) is illustrated. In FIG. 11, the base station (1120) may be a transmitting unit that transmits label data, and the terminal (1110) may be a receiving unit that receives a signal containing label data. However, as in FIG. 11, the learning of the artificial intelligence model may be performed on the transmitting unit side.

[0169] In the embodiment of FIG. 11, a base station (1120) transmits a signal containing label data to a terminal (1110), and the terminal (1110) can collect data for training an artificial intelligence model. Then, the terminal (1110) can transmit training data to the base station (1120) in accordance with the training data reporting instructions of the base station (1120), and the base station (1120) can train an artificial intelligence model using the received training data.

[0170] In step 1132, the base station (1120) may transmit information related to the generation of label data for training an artificial intelligence model to the terminal (1110). Based on the information related to the generation of label data, the terminal (1110) may identify the label data. At this time, the label data may represent bit information including at least one bit. That is, the label data may represent a vector including bits composed of 0 or 1. The information related to the generation of label data may include various information that enables the terminal (1110) to generate label data. For example, the information related to the generation of label data may include at least one of table information including multiple bit information and corresponding indices (see FIG. 8), table information including indices corresponding to sequence generation methods for generating label data (see FIG. 9), or information regarding a seed value for generating label data.

[0171] In step 1134, the base station (1120) can transmit resource information allocated for transmitting label data to the terminal (1110). That is, the base station (1120) can transmit scheduling information for transmitting label data to the terminal (1110). Accordingly, the terminal (1110) can identify time resources and frequency resources capable of receiving label data based on the received scheduling information.

[0172] In various embodiments, the base station (1120) may transmit to the terminal (1110) at least one of an index corresponding to specific bit information, an index corresponding to a specific sequence generation method, or information regarding a seed value for generating a specific sequence, according to information related to the generation of label data in step 1132. Additionally, the base station (1120) may transmit to the terminal (1110) parameters for generating label data (e.g., parameters applied to a mathematical formula for generating label data or parameters applied to a mathematical formula for generating a seed value). That is, the base station (1120) may provide information that enables the terminal (1110) to identify or calculate label data. Reference may be made to the content described in FIGS. 7 through 9 above regarding the information that enables the terminal (1110) to identify or calculate label data.

[0173] In step 1136, the base station (1120) may transmit a signal containing label data to the terminal (1110). The terminal (1110) may identify the label data or receive the signal containing the label data based on the information received in step 1132 and / or the information received in step 1134. For example, the terminal (1110) may identify the label data itself (e.g., bit information), a sequence for generating the label data, or a seed value for generating the label data based on the information received from the base station, and may know from the identified information what label data the signal received from the base station (1120) contains.

[0174] In various embodiments, the terminal (1110) may receive a signal containing label data from a base station (1120) based on resource scheduling information. At this time, the terminal (1110) may identify the location of a reference signal received from the base station based on the correlation of the reference signal to identify a signal corresponding to the label data, and may identify the location of a signal corresponding to the label data from the identified location of the reference signal.

[0175] The terminal (1110) compares the pattern of a received signal with the pattern of a reference signal, and if the correlation is greater than a specific value, identifies the signal pattern as a reference signal, and can identify the location of the signal corresponding to the label data based on the identified reference signal. Such an operation may be referred to as synchronization.

[0176] In various embodiments, the terminal (1110) can identify the location of a received reference signal based on the correlation between a reference signal pattern and a signal pattern corresponding to label data and a received signal pattern, and can identify the location of a signal corresponding to label data from the identified location of the reference signal. That is, when determining the correlation, the terminal (1110) can use the reference signal pattern and the signal pattern corresponding to label data together.

[0177] When verifying the correlation with the received signal by utilizing both the reference signal pattern and the signal pattern corresponding to the label data, it is possible to identify the location of the reference signal more accurately even in the presence of signal distortion and noise.

[0178] In various embodiments, when synchronization is ensured between the terminal (1110) and the base station (1120) based on TA (timing advance) information, the operation of identifying the location of the received reference signal based on correlation may be omitted.

[0179] In various embodiments, label data may be scheduled simultaneously with data containing other information (e.g., data not used for training an artificial intelligence model) (e.g., allocated to the same time resource), or each may be scheduled to a different time resource.

[0180] In step 1138, the terminal (1110) can obtain data for training an artificial intelligence model by processing a signal received from the base station (1120). For example, based on the data processing method of the receiving end described in FIG. 4, various data corresponding to input information for training an artificial intelligence model (e.g., You can obtain (etc.).

[0181] In addition, since the terminal receives information from the base station for identifying or generating label data, it can identify or generate label data. That is, the terminal can configure label data and data to be input into an artificial intelligence model as data for training the artificial intelligence model.

[0182] In step 1140, the base station (1120) may instruct the terminal (1110) to report data for training an artificial intelligence model.

[0183] In various embodiments, the reporting instruction of step 1140 may instruct the terminal (1110) to report label data, or may instruct it not to report label data. That is, it may instruct whether or not to report label data.

[0184] Since the base station (1120) is the entity that transmitted the label data, it can know the label data corresponding to the input data for artificial intelligence model training reported by the terminal (1110), and in this case, it can instruct the terminal (1110) not to report the label data.

[0185] In various embodiments, the action of instructing to report data may represent an action of instructing a separate test terminal or an action of instructing a terminal for actual use.

[0186] In various embodiments, the action of instructing to report data may represent an action of instructing to transmit data of a specific size at a specific time. For example, it may represent instructing to transmit data for 10 slots at the top of every hour, or instructing to transmit data during the evening hours when data traffic is low.

[0187] In various embodiments, the operation of instructing to report data may selectively instruct reporting by taking into account the mobility or location information of the terminal.

[0188] In step 1142, the terminal (1110) can transmit data for artificial intelligence model training to the base station (1120) based on the report instructions from the base station (1120) received in step 1140. Based on the report instructions from the base station (1120), the terminal (1110) may or may not include label data in the data for artificial intelligence model training. Additionally, the terminal may prioritize the accuracy of the transmitted information based on a method of adjusting the modulation coding scheme (MCS) to minimize signal distortion and noise caused by retransmitting the data for artificial intelligence model training.

[0189] In step 1144, the base station (1120) can train an artificial intelligence model based on the artificial intelligence model training data received from the terminal in step 1142 described above. The artificial intelligence model training data can be configured in various ways depending on the type of artificial intelligence model, and the type of artificial intelligence model can be distinguished based on which processes of the received signal processing procedure the model can replace.

[0190] For example, in the case of an artificial intelligence model to replace the processes of soft demodulation (4212), descrambling (4214), rate dematching (4216), and channel decoding (4218) as shown in FIG. 4, the data for training the artificial intelligence model is the data generated from the channel equalizer as input data. Data that may include, and includes a bit value of 0 or 1 as label data It may include.

[0191] Additionally, for example, in the case of an artificial intelligence model to replace the operation of a channel equalizer (4210) and soft demodulation (4212) as shown in FIG. 6, the data for training the artificial intelligence model is the layer and RE demodulated data as input data. It may include, and data as label data It may include.

[0192] In various embodiments, the label data is bit data composed of bit values ​​of 0 or 1. Data generated by processing from (e.g., It may include (etc.). The terminal or base station may configure the required label data according to the type of artificial intelligence model to be learned.

[0193] In various embodiments, the input data and label data for training the artificial intelligence model may be configured differently depending on the type of artificial intelligence model.

[0194] FIG. 12 illustrates a data transmission method for learning an artificial intelligence model according to embodiments of the present disclosure.

[0195] Referring to FIG. 12, a transmission and reception operation between a terminal (1210) and a base station (1220) is illustrated. In FIG. 12, the terminal (1210) may be a receiving unit that receives label data, and the base station (1220) may be a transmitting unit that transmits a signal containing label data. An embodiment of FIG. 12 illustrates an embodiment in which learning of an artificial intelligence model is performed at the terminal (1210).

[0196] In step 1232, the base station (1220) may transmit information related to the generation of label data for training an artificial intelligence model to the terminal (1210). Based on the information related to the generation of label data, the terminal (1210) may generate label data. At this time, the label data may represent bit information including at least one bit. That is, the label data may represent a vector including bits composed of 0 or 1. The information related to the generation of label data may include various information that enables the terminal (1210) to generate label data by the instructions of the base station (1220). For example, the information related to the generation of label data may include at least one of table information including multiple bit information and corresponding indices (see FIG. 8), table information including indices corresponding to sequence generation methods for generating label data (see FIG. 9), or information regarding a seed value for generating label data.

[0197] In step 1234, the base station (1220) can transmit resource information allocated to transmit label data to the terminal (1210). That is, the base station (1220) can transmit scheduling information for transmitting label data to the terminal (1210). Accordingly, the terminal (1210) can identify time resources and frequency resources capable of receiving label data based on the received scheduling information.

[0198] In various embodiments, the base station (1220) may transmit to the terminal (1210) at least one of an index corresponding to specific bit information, an index corresponding to a specific sequence generation method, or information regarding a seed value for generating a specific sequence, according to information related to the generation of label data in step 1232. Additionally, the base station (1220) may transmit to the terminal (1210) parameters for generating label data (e.g., parameters applied to a mathematical formula for generating label data or parameters applied to a mathematical formula for generating a seed value). That is, the base station (1220) may provide information that enables the terminal (1210) to identify or calculate label data. Reference may be made to the content described in FIGS. 7 through 9 above regarding the information that enables the terminal (1210) to identify or calculate label data.

[0199] In step 1236, the base station (1220) may transmit a signal containing label data to the terminal (1210). The terminal (1210) may identify the label data or receive the signal containing the label data based on the information received in step 1232 and / or the information received in step 1234. For example, the terminal (1210) may identify the label data itself (e.g., bit information), a sequence for generating the label data, or a seed value for generating the label data based on the information received from the base station, and may know from the identified information what label data the signal received from the base station (1220) contains.

[0200] Likewise, in step 1236, the terminal (1210) may receive a signal containing label data from the base station (1220). At this time, the terminal (1210) may identify the location of the reference signal received from the terminal based on the correlation of the reference signal to identify the signal corresponding to the label data, and may identify the location of the signal corresponding to the label data from the identified location of the reference signal. The terminal (1210) may compare the pattern of the received signal with the pattern of the reference signal, and if the correlation is greater than or equal to a specific value, identify the corresponding signal pattern as the reference signal, and identify the location of the signal corresponding to the label data based on the identified reference signal. Such an operation may be referred to as synchronization.

[0201] In various embodiments, the terminal (1210) can identify the location of a received reference signal based on the correlation between a reference signal pattern and a signal pattern corresponding to label data and a received signal pattern, and can identify the location of a signal corresponding to label data from the identified location of the reference signal. That is, the terminal (1210) can use the reference signal pattern and the signal pattern corresponding to label data together when determining the correlation. When confirming the correlation with the received signal by using both the reference signal pattern and the signal pattern corresponding to label data, there is an effect of being able to identify the location of the reference signal more accurately even if there is signal distortion and noise.

[0202] In various embodiments, when synchronization is ensured between the terminal (1210) and the base station (1220) based on TA (timing advance) information, the operation of identifying the location of the received reference signal based on correlation may be omitted.

[0203] In step 1238, the terminal (1210) can process a signal received from the base station (1220) to obtain data for training an artificial intelligence model. For example, based on the data processing method of the receiving end described in FIG. 4, various data corresponding to input information for training an artificial intelligence model (e.g., You can obtain (etc.).

[0204] In step 1240, the terminal (1210) can train an artificial intelligence model based on the artificial intelligence model training data collected in the aforementioned step 1238. The artificial intelligence model training data can be configured in various ways according to the received data processing procedure to be replaced by the artificial intelligence model. For example, in the case of an artificial intelligence model to replace soft demodulation (4212), descrambling (4214), rate dematching (4216), and channel decoding (4218) operations as shown in FIG. 4, the artificial intelligence model training data is data generated from a channel equalizer as input data. Data that may include, and includes a bit value of 0 or 1 as label data It may include. Additionally, for example, in the case of an artificial intelligence model to replace the operation of a channel equalizer (4210) and soft demodulation (4212) as in FIG. 6, the data for training the artificial intelligence model may include layer and RE-demapped data as input data. It may include, and data as label data It may include.

[0205] In various embodiments, the label data is bit data composed of bit values ​​of 0 or 1. Data generated by processing from (e.g., It may include (etc.). The terminal (1210) or base station (1220) may configure the necessary label data according to the type of artificial intelligence model.

[0206] In various embodiments, the input data and label data for training the artificial intelligence model may be configured differently depending on the type of artificial intelligence model.

[0207] FIG. 13 shows input data and output data for training an artificial intelligence model according to embodiments of the present disclosure.

[0208] Referring to FIG. 13, a method for training an artificial intelligence model (hereinafter, the first model) to replace the channel equalizer, soft demodulation, descrambling, rate dematching, and channel decoding processes in relation to the processing procedure of a received signal is illustrated.

[0209] The first model receives channel information estimated by the channel estimator and / or output information of the layer and RE demapping operation as input data, and the restored bit information It can output. The input to the first model may further include the number of layers, the modulation coding scheme (MCS), and / or the signal-to-noise ratio (SNR), and such information corresponds to information known to the receiver or transmitter for communication. Recovered bit information output from the first model is the original bit information transmitted from the transmitting end. True or false can be determined by (or, label data), and the learning process of the first model is the first model is the restored bit information It can proceed in a direction that increases the accuracy of.

[0210] In various embodiments, the operation of collecting data for training an artificial intelligence model may represent the operation of collecting input data necessary for training the model, such as channel information, output information of layer and RE demapping operations, SNR, MCS and / or label data, and label data for determining the output value of the model being trained.

[0211] FIG. 14 shows input data and output data for training an artificial intelligence model according to embodiments of the present disclosure.

[0212] Referring to FIG. 14, a method for training an artificial intelligence model (hereinafter referred to as the second model) to replace the processing process of a channel equalizer and a channel estimator in relation to the processing procedure of a received signal is illustrated. The second model can determine the receiving mode and perform equalizing operations.

[0213] The second model can receive reference signals such as DMRS and SRS and / or output information of layers and RE demapping operations as input data and output a receive mode (e.g., RX type). The input to the second model may further include the number of layers, modulation coding scheme (MCS), and / or signal-to-noise ratio (SNR), and such information corresponds to information known to the receiver or transmitter for communication.

[0214] In various embodiments, the receiver may derive a block error rate, a bit error rate, and / or a symbol error rate based on data restored from a received signal. Accordingly, the receiver may use the complexity, block error rate (BLER), bit error rate (BER), and / or symbol error rate according to the reception mode as training data (input data or label data) for the second model.

[0215] In various embodiments, the second model may output a block error rate, a bit error rate, and / or a symbol error rate according to the receiving mode, or output a selected receiving mode based thereon. Additionally, the second model recovers data It can output, It can be processed based on the receiving mode output from the second model.

[0216] In various embodiments, the receiving mode may include ZF (zero forcing), MMSE (minimum mean squared error), MRC (maximal ratio combining), etc.

[0217] In various embodiments, the operation of collecting data for training an artificial intelligence model may represent the operation of collecting input data necessary for training the model, such as channel information, output information of layer and RE demapping operations, SNR, MCS and / or label data, and label data for determining the output value of the model being trained.

[0218] A method performed by a first electronic device according to various embodiments of the present disclosure may include: transmitting information to a second electronic device for generating label data including at least one bit value; instructing the second electronic device to transmit the label data; receiving a signal including the label data from the second electronic device; processing the received signal to obtain input data for training an artificial intelligence model; and training an artificial intelligence model based on the input data and the label data.

[0219] In various embodiments, the first electronic device may include a receiving end, and the second electronic device may include a transmitting end. The first electronic device and the second electronic device may each represent a terminal or a base station.

[0220] In various embodiments, information for generating label data may include at least one of first information (see FIG. 8) comprising a plurality of indexes corresponding to a plurality of label data, second information (see FIG. 9) comprising a plurality of indexes corresponding to a plurality of sequence generation methods, or third information regarding a randomly determined seed value. The third information may include information regarding a seed value, a mathematical formula for generating the seed value, and / or parameters applied to the mathematical formula for generating the seed value. Additionally, the mathematical formula for generating the seed value may be based on the number of symbols per slot, a slot number within a frame, an OFDM symbol number, a scrambling identifier, or a physical layer cell identifier.

[0221] In various embodiments, the step of directing the transmission of label data may be performed based on RRC (radio resource control) signaling, MAC (medium access control) CE (control element) signaling, or DCI (downlink control indicator) signaling.

[0222] Additionally, the step of the first electronic device directing the transmission of label data may include the step of transmitting at least one of an index directing the label data, an index directing a sequence generation method for generating the label data, or a seed value for generating the label data.

[0223] In various embodiments, the step of a first electronic device directing the transmission of label data further includes the step of allocating time resources and frequency resources for the transmission of label data, and the step of transmitting information about the time resources and frequency resources to a second electronic device, and the step of receiving a signal including label data can receive the signal based on the time resources and frequency resources.

[0224] In various embodiments, the step of a first electronic device receiving a signal containing label data may include identifying the location of the reference signal from the received signal based on the correlation between the reference signal and the received signal, and identifying the location of a signal corresponding to the label data from the received signal based on the location of the reference signal.

[0225] In various embodiments, the step of the first electronic device identifying the location of a reference signal can identify the location of the reference signal from the received signal based on the correlation between the signal corresponding to the reference signal and the label data and the received signal.

[0226] According to the embodiments of the present disclosure, there is an effect that even received signals that cannot be processed at the receiving end can be utilized as data for training an artificial intelligence model. Therefore, since the data for training the artificial intelligence model is not dependent on the signal processing performance of the receiving end, the performance of the artificial intelligence model can be improved.

[0227] Although the embodiments of the present disclosure have been described in detail to explain the technical concept of the present disclosure, the individual operations constituting each embodiment may be changed in order or parts of which may be omitted. Accordingly, an embodiment in which the order of operations is changed or some operations are omitted may be understood as having been described by the present disclosure. Furthermore, the embodiments of the present disclosure may be modified in various ways according to the content described in the present disclosure.

[0228] Although the operations of the method according to the embodiments of the present disclosure have been described separately for each embodiment, the operations included in each embodiment may be combined with the operations of other embodiments to form new embodiments. Accordingly, embodiments in which the embodiments of the present disclosure are combined may also be understood as being described by the present disclosure.

[0229] The various embodiments of the present disclosure and the terms used therein are not intended to limit the technical features described in the present disclosure to specific embodiments, and should be understood to include various modifications, equivalents, or substitutions of said embodiments. In connection with the description of the drawings, similar reference numerals may be used for similar or related components. The singular form of a noun corresponding to an item may include one or more items unless the relevant context clearly indicates otherwise. In the present disclosure, each of the phrases such as “A or B”, “at least one of A and B”, “at least one of A or B”, “A, B or C”, “at least one of A, B and C”, and “at least one of A, B, or C” may include any one of the items listed together in the corresponding phrase, or all possible combinations thereof. Terms such as "first," "second," or "first" or "second" may be used simply to distinguish a component from another component and do not limit the components in any other aspect (e.g., importance or order). Where a component (e.g., the first) is referred to as "coupled" or "connected" to another component (e.g., the second), with or without the terms "functionally" or "communicationally," it means that the component may be connected to the other component directly (e.g., via a wire), wirelessly, or through a third component.

[0230] As used in this disclosure, the term "module" may include a unit implemented in hardware, software, or firmware, and may be used interchangeably with terms such as logic, logic block, component, or circuit. A module may be a component formed integrally, or a minimum unit of a component or part thereof that performs one or more functions. For example, according to one embodiment, a module may be implemented in the form of an application-specific integrated circuit (ASIC).

[0231] Various embodiments of the present disclosure may be implemented as software (e.g., a program) comprising one or more instructions stored in a storage medium (e.g., internal memory or external memory) readable by a machine (e.g., an electronic device). For example, a machine (e.g., a processor of an electronic device (e.g., processor (230)) may call at least one of the one or more instructions stored from the storage medium and execute it. This enables the machine to operate to perform at least one function according to at least one called instruction. One or more instructions may include code generated by a compiler or code that can be executed by an interpreter. The storage medium readable by the machine may be provided in the form of a non-transitory storage medium. Here, "non-transitory" simply means that the storage medium is a tangible device and does not contain a signal (e.g., an EM wave), and this term does not distinguish between cases where data is stored semi-permanently and cases where it is stored temporarily in the storage medium.

[0232] According to one embodiment, the method according to the various embodiments disclosed herein may be provided by being included in a computer program product. The computer program product may be traded between a seller and a buyer as a product. The computer program product may be distributed in the form of a device-readable storage medium (e.g., CD-ROM (compact disc read-only memory)), or distributed online (e.g., download or upload) through an application store (e.g., Play Store™) or directly between two user devices (e.g., smartphones). In the case of online distribution, at least a portion of the computer program product may be temporarily stored or temporarily created in a device-readable storage medium, such as the memory of a manufacturer's server, an application store's server, or a relay server.

[0233] According to various embodiments, each component (e.g., module or program) of the described components may include a singular or multiple entities. According to various embodiments, one or more of the components or operations among the aforementioned components may be omitted, or one or more other components or operations may be added. Generally or additionally, multiple components (e.g., module or program) may be integrated into a single component. In this case, the integrated component may perform one or more functions of each of the multiple components in the same or similar manner as they were performed by the corresponding component among the multiple components prior to integration. According to various embodiments, operations performed by a module, program, or other component may be executed sequentially, in parallel, iteratively, or heuristically; one or more of the operations may be executed in a different order; omitted; or one or more other operations may be added.

Claims

1. A method performed by a first electronic device in a wireless communication system, A step of transmitting information to a second electronic device for generating label data including at least one bit value; A step of transmitting a transmission instruction for the label data to the second electronic device; A step of receiving a signal including the label data from the second electronic device; A step of processing the received signal to obtain input data for training an artificial intelligence model; and A step comprising training the artificial intelligence model based on the input data and label data. method.

2. In Claim 1, The information for generating the above label data is, It includes at least one of first information comprising a plurality of label data and a plurality of corresponding indices, second information comprising a plurality of sequence generation methods and a plurality of corresponding indices, or third information regarding a randomly determined seed value, and The step of instructing the transmission of the above label data is, Performed based on RRC (radio resource control) signaling, MAC (medium access control) CE (control element) signaling, or DCI (downlink control indicator) signaling, method.

3. In Claim 2, The above third information is, It includes at least one of information regarding the seed value or a formula for generating the seed value, and The above formula is, Based on the number of symbols per slot, the slot number within the frame, the OFDM symbol number, the scrambling identifier, or the physical layer cell identifier, method.

4. In Claim 3, The step of instructing the transmission of the above label data is, A method comprising the step of transmitting at least one of an index indicating the label data, an index indicating a sequence generation method for generating the label data, or a seed value for generating the label data. method.

5. In Claim 3, The step of instructing the transmission of the above label data is, The step of allocating time resources and frequency resources for transmitting the above label data, and The method further includes the step of transmitting information regarding the time resource and the frequency resource to the second electronic device. The step of receiving a signal including the label data above involves receiving the signal based on the time resource and the frequency resource, method.

6. In Claim 1, The step of receiving a signal including the above-mentioned label data is, A step of identifying the location of the reference signal from the received signal based on the correlation between the reference signal and the received signal, and Based on the location of the reference signal, the method includes the step of identifying the location of a signal corresponding to the label data from the received signal. The step of identifying the location of the above reference signal is, Identifying the location of the reference signal from the received signal based on the correlation between the signal corresponding to the reference signal and the label data, method.

7. In Claim 1, The first electronic device above includes a base station, and The second electronic device above includes a user terminal (UE), method.

8. A method performed by a second electronic device in a wireless communication system, A step of receiving information from a first electronic device for generating label data including at least one bit value; A step of receiving a transmission instruction for the label data from the first electronic device; and The method comprises the step of transmitting a signal including the label data to the first electronic device. method.

9. As a first electronic device of a wireless communication system, At least one transceiver; At least one processor communicatively coupled to the above at least one transceiver; and It includes at least one memory that is communicationally coupled to the above at least one processor and stores instructions, and The above instructions are executed individually or in any combination by the at least one processor, so that the first electronic device: Information for generating label data including at least one bit value is transmitted to a second electronic device, and Transmitting a transmission instruction for the label data to the second electronic device, and A signal including the label data is received from the second electronic device, and Process the received signal above to obtain input data for training an artificial intelligence model, and Training the artificial intelligence model based on the above input data and the above label data, First electronic device.

10. In Claim 9, The information for generating the above label data is, It includes at least one of first information comprising a plurality of label data and a plurality of corresponding indices, second information comprising a plurality of sequence generation methods and a plurality of corresponding indices, or third information regarding a randomly determined seed value. The transmission instruction for the above label data is transmitted based on RRC (radio resource control) signaling, MAC (medium access control) CE (control element) signaling, or DCI (downlink control indicator) signaling, First electronic device.

11. In Claim 10, The above third information is, It includes at least one of information regarding the seed value or a formula for generating the seed value, and The above formula is, Based on the number of symbols per slot, the slot number within the frame, the OFDM symbol number, the scrambling identifier, or the physical layer cell identifier, First electronic device.

12. In Claim 11, The above commands are the first electronic device: Transmitting a transmission instruction for the label data based on at least one of an index indicating the label data, an index indicating a sequence generation method for generating the label data, or a seed value for generating the label data. First electronic device.

13. In Claim 11, The above commands are the first electronic device: Allocate time and frequency resources for the transmission of the above label data, and Transmitting information regarding the time resource and the frequency resource to the second electronic device, Receiving a signal including the label data based on the above time resources and the above frequency resources, First electronic device.

14. In Claim 11, The above commands are the first electronic device: Based on the correlation between the reference signal and the received signal, the location of the reference signal is identified from the received signal, and Based on the location of the reference signal above, the location of the signal corresponding to the label data is identified from the received signal, and Based on the correlation between the signal corresponding to the reference signal and the label data and the received signal, the method for identifying the location of the reference signal from the received signal. First electronic device.

15. As a second electronic device of a wireless communication system, At least one transceiver; At least one processor communicatively coupled to the above at least one transceiver; and It includes at least one memory that is communicationally coupled to the above at least one processor and stores instructions, and The above instructions are executed individually or in any combination by the at least one processor, so that the second electronic device: Information for generating label data including at least one bit value is received from a first electronic device, and Receiving a transmission instruction for the label data from the first electronic device, and The first electronic device is configured to transmit a signal including the label data. Second electronic device.

Images