Electronic device for estimating speed of terminal on basis of artificial intelligence, and operation method thereof

Abstract

Provided, according to one embodiment, may be a method of a base station comprising at least one AI-based inference unit in a wireless communication system. The method may comprise an operation of receiving, from a user equipment (UE), channel state information (CSI) including a precoding matrix indicator (PMI), a channel quality indicator (CQI), and a rank indicator (RI). The method may comprise an operation of performing a normalization procedure to adjust respective values of the PMI, the CQI, and the RI within preconfigured ranges. The method may comprise an operation of determining, via at least one inference unit included in the base station, whether the speed of the UE is high or low by using the normalized PMI, CQI, and RI. The method may comprise an operation of allocating resources to the UE according to whether the speed of the UE is high or low.

Description

Electronic device for estimating the speed of a terminal based on artificial intelligence and its operating method

[0001] The present disclosure relates to an electronic device for estimating the speed of a UE based on artificial intelligence and an operating method thereof.

[0002] With the development of digital technology, electronic devices are available in various forms, such as smart phones, tablet personal computers, and personal digital assistants (PDAs).

[0003] As artificial intelligence (AI) technology advances, electronic devices can apply AI technology to provide a variety of AI services. Electronic devices can provide AI services by providing software components (e.g., services, functions, or programs) that process user-requested tasks and provide customized services (e.g., customized information based on user voice commands) based on AI and voice recognition technologies.

[0004] Machine learning is a field related to artificial intelligence, developing algorithms and technologies that enable computers to learn. Deep learning refers to a set of machine learning algorithms that attempt to achieve a high level of abstraction (the task of extracting only the essential information from large amounts of complex data) through a combination of nonlinear transformation methods.

[0005] The above information may be provided as background information to aid in understanding this document. None of the above is claimed to be prior art related to this document or can be used to determine prior art.

[0006] The present disclosure proposes a method for estimating the speed of a UE based on artificial intelligence.

[0007] According to one embodiment, a method of a base station including at least one AI-based inference device in a wireless communication system may be provided. The method may include receiving channel state information (CSI) including a precoding matrix indicator (PMI), a channel quality indicator (CQI), and a rank indicator (RI) from a user equipment (UE). The method may include performing a normalization procedure to adjust the values ​​of each of the PMI, the CQI, and the RI to within a preset range. The method may include determining, through at least one inference unit included in the base station, whether a speed of the UE is high or low using the normalized PMI, CQI, and RI. The method may include allocating resources to the UE based on whether the speed of the UE is high or low.

[0008] According to one embodiment, a storage medium storing at least one computer-readable instruction may be provided. The at least one instruction, when executed by at least a part of at least one processor of an electronic device, may cause the electronic device to perform at least one operation. The at least one operation may include receiving channel state information (CSI) including a precoding matrix indicator (PMI), a channel quality indicator (CQI), and a rank indicator (RI) from a user equipment (UE). The at least one operation may include performing a normalization procedure to adjust the values ​​of each of the PMI, the CQI, and the RI within a preset range. The at least one operation may include determining, through at least one inference unit included in the base station, whether a speed of the UE is high or low using the normalized PMI, CQI, and RI. The at least one operation may include allocating resources to the UE based on whether the speed of the UE is high or low.

[0009] According to one embodiment, an electronic device may include at least one processor. The electronic device may include a memory storing at least one instruction. The at least one instruction, when executed by at least a portion of the at least one processor, may cause the electronic device to perform at least one operation. The at least one operation may include receiving channel state information (CSI) including a precoding matrix indicator (PMI), a channel quality indicator (CQI), and a rank indicator (RI) from a user equipment (UE). The at least one operation may include performing a normalization procedure to adjust the values ​​of each of the PMI, the CQI, and the RI to within a preset range. The at least one operation may include determining, through at least one inference unit included in the base station, whether a speed of the UE is high or low using the normalized PMI, CQI, and RI. The at least one operation may include an operation of allocating resources to the UE depending on whether the UE has a high speed or a low speed.

[0010] The method and device according to the embodiment of the present disclosure can determine whether the speed of a UE is high or low using an AI-based inference device, and efficiently allocate resources to the UE according to the determined speed of the UE.

[0011] In connection with the description of the drawings, the same or similar reference numerals may be used for the same or similar components.

[0012] FIG. 1 illustrates a block diagram of an electronic device according to one embodiment of the present disclosure.

[0013] FIG. 2 illustrates a block diagram of an electronic device according to one embodiment of the present disclosure.

[0014] FIG. 3 illustrates an example of assigning different DMRS types depending on the speed of a UE according to one embodiment of the present disclosure.

[0015] FIG. 4 is a diagram for explaining the operation of a base station according to one embodiment of the present disclosure.

[0016] FIG. 5 illustrates an example of a transmission and reception procedure between a base station and a UE for determining a speed of the UE according to one embodiment of the present disclosure.

[0017] FIG. 6 is a diagram for explaining the operation of an AI-based UE speed classifier according to one embodiment of the present disclosure.

[0018] FIG. 7 illustrates an example of a base station including multiple AI-based UE speed classifiers according to one embodiment of the present disclosure.

[0019] FIG. 8 illustrates an example of a distorted signal having a normal distribution according to one embodiment of the present disclosure.

[0020] FIG. 9 illustrates an example of a change in an input signal after signal distortion according to one embodiment of the present disclosure.

[0021] FIG. 10 illustrates an example of an inference result of an AI model according to one embodiment of the present disclosure.

[0022] FIG. 11 is a diagram for explaining the operation of a base station according to one embodiment of the present disclosure.

[0023] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings so that those skilled in the art can easily implement the present disclosure. However, the present disclosure may be implemented in various different forms and is not limited to the embodiments described herein. In connection with the description of the drawings, the same or similar reference numerals may be used for identical or similar components. Furthermore, in the drawings and related descriptions, descriptions of well-known functions and configurations may be omitted for clarity and conciseness.

[0024] FIG. 1 illustrates a block diagram of an electronic device according to one embodiment of the present disclosure.

[0025] Referring to FIG. 1, an electronic device (101) may include a processor (120), a memory (130), and / or a communication interface (190). The electronic device (101) may be a device for providing a service linked to at least one AI model. The electronic device (101) may be implemented as, for example, a single entity, or may be implemented as multiple entities. For example, the electronic device (101) may be implemented as a base station, a user equipment, or a server, and there is no limitation on the form of implementation.

[0026] The processor (120) may, for example, execute software (e.g., a program) to control at least one other component (e.g., a hardware or software component) of the electronic device (101) connected to the processor (120) and perform various data processing or operations. According to one embodiment, as at least a part of the data processing or operations, the processor (120) may store commands or data received from another component (e.g., a communication interface (190)) in a memory (130) (e.g., a volatile memory, but not limited thereto), process the commands or data stored in the memory (130), and store resultant data in the memory (130) (e.g., a non-volatile memory, but not limited thereto). According to one embodiment, the processor (120) may include, but is not limited to, a central processing unit (CPU), a GPU, and / or a neural processing unit (NPU) including circuitry.

[0027] The memory (130) may store various data used by at least one component (e.g., the processor (120) and / or the communication interface (190)) of the electronic device (101). The data may include, for example, software (e.g., a program) and input data or output data for commands related thereto. The memory (130) may include volatile memory and / or non-volatile memory. The memory (130) may include a hard disk, ROM, RAM, cache memory, and / or registers, and the implementation thereof is not limited thereto. Some of the above-described entities (e.g., registers, but not limited thereto) may be implemented as part of the processor (120), and the implementation form thereof is not limited thereto. The memory (130) may also store at least one AI model for determining the properties of the electronic device (101) or an external electronic device.

[0028] The processor (120) may execute, for example, at least one instruction stored in the memory (130). The memory (130) may store at least one instruction, and the at least one instruction may be executed by the processor (120). The at least one instruction, when executed by the processor (120), may cause the electronic device (101) to perform at least one operation. For example, as the at least one instruction is executed, at least one other component may be controlled, and / or various data processing or operations may be performed. As at least a part of the data processing or operations, the processor (120) may store commands or data received from other components in at least a part of the memory (130), process the commands or data stored in the memory (130), and store result data in the memory (130). A processor (120) performing an operation may mean, for example, that the operation is performed by (or under the control of) one entity included in the processor (120) (e.g., but not limited to, the main processor). For example, a processor performing an operation may mean, for example, that a specific operation is performed by (or under the control of) multiple entities (e.g., multiple processors). For example, a plurality of operations may mean, for example, that all of the plurality of operations are performed by (or under the control of) one entity (e.g., but not limited to, the main processor). For example, a plurality of operations may mean that some of the plurality of operations are performed by at least one entity, and some of the remaining operations are performed by at least one other entity. For example, at least one instruction causing the performance of one or more operations may be stored in one memory, or may be stored in a distributed manner in each of a plurality of memories.

[0029] The communication interface (190) may support the establishment of a direct (e.g., wired) communication channel or a wireless communication channel between the electronic device (101) and an external electronic device, and the performance of communication through the established communication channel. The communication interface (190) may include one or more communication processors that operate independently from the processor (120) and support direct (e.g., wired) communication or wireless communication, but is not limited thereto. According to one embodiment, the communication interface (190) may include an Ethernet-based wired communication module (e.g., a local area network (LAN) communication module or a power line communication module). The communication interface (190) may also include a wireless communication module or a transceiver, but is not limited thereto. At least some of the above components may be connected to each other through a communication method between peripheral devices (e.g., a bus, a general purpose input and output (GPIO), a serial peripheral interface (SPI), or a mobile industry processor interface (MIPI)) and may exchange signals (e.g., commands or data) with each other.

[0030] FIG. 2 illustrates a block diagram of an electronic device according to one embodiment of the present disclosure.

[0031] According to one embodiment, the electronic device (101) may include and / or execute a frontend module (210) and / or a service processing module (220). The frontend module (210) and / or the service processing module (220) may be executed by, for example, the processor (120), or may be included as at least a part of the processor (120) or another entity. At least some of the operations performed by the frontend module (210) and / or the service processing module (220) in the present disclosure may be understood to be performed by, for example, the processor (120) and / or another entity under the control of the processor (120).

[0032] The front-end module (210) can perform at least one operation for data exchange with, for example, external electronic devices (106a, 106b, 106c, ..., 106n). For example, the front-end module (210) can provide data that can configure a user interface (UI) that can input user input from the external electronic devices (106a, 106b, 106c, ..., 106n). The front-end module (210) can provide processing for a user request to the service processing module (220). The service processing module (220) can perform a service using the user request and may also be referred to as a back-end module. The service processing module (220) can provide a response corresponding to the user request to the front-end module (210). The front-end module (210) can provide a response received from the service processing module (220) to an external electronic device (106a, 106b, 106c, ..., 106n).

[0033] The service processing module (220) may include, for example, a user request confirmation module (221), an optimization module (222), an AI model management module (224), a policy management module (226), and / or a service execution module (227).

[0034] The user request verification module (221) can verify information associated with a user request provided from an external electronic device (106a, 106b, 106c, ..., 106n). The information associated with the user request can be expressed as, for example, the number of user requests over a certain period of time and / or the size of the user request, but there is no limitation thereto. For example, the user request verification module (221) can count the number of user requests provided from the external electronic device (106a, 106b, 106c, ..., 106n) and / or monitor the size of the content included in the user request (for example, text and / or graphic objects, but there is no limitation thereto), but there is no limitation on the type of information associated with the user request and / or the verification method.

[0035] The optimization module (222) can provide at least one optimal number of instances for each of at least one AI model linked to the service.

[0036] The AI ​​model management module (224) can store and / or manage (e.g., including but not limited to, adding, deleting, and / or updating) AI models associated with a service.

[0037] The policy management module (226) may, for example, store and / or manage (e.g., including but not limited to, adding, deleting, and / or updating) acceptable response times. For example, the management device (104) may verify (e.g., receive or determine) at least one input for determining an acceptable response time and provide it to the electronic device (101). For example, an administrator may input information regarding an acceptable response time for a given service into the management device (104), but this is exemplary and there is no limitation on the manner in which the acceptable response time is verified.

[0038] The service execution module (227) may execute, for example, at least one instance group (231, 232, 233) corresponding to at least one AI model linked to the service. The service execution module (227) may execute, for example, at least one instance group (231, 232, 233) corresponding to at least one AI model according to the optimal number of instances provided by the optimization module (222). The service execution module (227) may process a user request based on the executed at least one instance group (231, 232, 233) and provide a response according to the processing result to an external electronic device (106a, 106b, 106c, ..., 106n) through the front-end module (210). According to one embodiment, when there are multiple instances being executed, the user request may be distributedly processed by the multiple instances.

[0039] Recently, in order to meet the diverse implementation requirements of various wireless services such as the NR (New Radio) standard, which is a 5th generation data transmission method, enhanced Mobile Broad-Band (eMBB), Ultra Reliable Low Latency Communication (URLLC), and / or the Internet-of-Things (IoT), various attempts are being made to improve the quality of service provided by base stations through speed estimation of user equipment (UE). For example, when the UE is in a low-speed environment, the base station can reduce the overhead of control channels other than data transmission in order to provide large-capacity services. For example, the base station can adjust the Modulation and Coding System (MCS) level and / or the number of resource blocks (RBs) after determining the UE's speed.

[0040] Meanwhile, when estimating the UE's speed using the Sounding Reference Signal (SRS) value transmitted by the UE as input, the base station must store multiple SRS measurement values, which may increase the demand for memory space. Furthermore, since the SRS values ​​are composed of complex real numbers, the I / O burden transferred from memory to the base station's inference unit may also increase, and the computational complexity may increase depending on the input features of the inference.

[0041] Meanwhile, when estimating the UE's speed using the CSI (channel state information) report transmitted by the UE, the I / O is simpler than with the SRS technique, but the CSI prediction result can be used when the base station infers the UE's speed. The CSI prediction result can be viewed as a preprocessing step for UE speed estimation, which can incur additional complexity. Even if CSI prediction is performed using a simple algorithm, inaccurate CSI prediction can deteriorate the UE's speed estimation accuracy.

[0042] In this disclosure, we propose a UE speed estimation method and structure that utilizes CSI report information transmitted by a UE to a base station for better speed estimation. According to one embodiment, the base station can determine whether the UE is at a low speed or a high speed through an AI model that inputs at least one of a precoding matrix indicator (PMI), a channel quality indicator (CQI), and a rank indicator (RI) included in the CSI report information. By using at least one of the PMI, CQI, and RI values ​​when determining the UE speed, the base station can reduce the memory required to store input data required for inference compared to an SRS-based UE speed determination technique (because the PMI, CQI, and RI values ​​can be expressed with a smaller number of bits than the SRS). By using at least one of the PMI, CQI, and RI values ​​when determining the UE speed, the base station can simplify the input of the AI ​​model for determining the UE speed, thereby securing an advantage in terms of complexity. The above base station uses at least one of the PMI, CQI, and RI values ​​when determining the UE speed, so there is a low possibility of performance degradation due to inaccurate preprocessing, and it can also be convenient in terms of model management because only one AI model is used.

[0043] FIG. 3 illustrates an example of allocating different DMRS (Demodulation Reference Signal) types according to the speed of a UE according to one embodiment of the present disclosure.

[0044] PDSCH (physical downlink shared channel) DMRS can be a special type of physical layer signal that serves as a reference signal for PDSCH decoding. LTE does not require this special DMRS for PDSCH because it can use the Cell Specific Reference Signal (CRS) for PDSCH decoding. However, 5G / NR may require a DMRS dedicated to PDSCH decoding. The base station transmits the DMRS for PDSCH decoding to the UE, and the UE can receive data on the PDSCH based on the DMRS.

[0045] The base station can determine the speed of the UE (e.g., estimate whether the UE is at a low / high speed) through an AI model that takes as input at least one of the PMI, CQI, and RI values ​​included in the CSI report transmitted by the UE to the base station. The speed characteristics of the UE inferred (or estimated) through the AI ​​model within the base station can be transmitted to the scheduler within the base station and utilized for resource allocation for the UE.

[0046] Referring to FIG. 3, a resource for DMRS allocation may be composed of at least one subcarrier included in a resource block in a frequency domain, and at least one OFDM symbol included in a slot in a time domain. One subcarrier in the frequency domain and one symbol in the time domain may constitute a resource element (RE). According to one embodiment, the minimum RE group allocated to the frequency domain in DMRS type 1 may be one RE. According to one embodiment, the minimum RE group allocated to the frequency domain in DMRS type 2 may be two consecutive REs.

[0047] If the base station determines that a UE is a high-speed UE through the AI ​​model, it can allocate DMRS type 1 resources to the high-speed UE for accurate channel estimation. If the base station determines that a UE is a low-speed UE through the AI ​​model, it can allocate DMRS type 2 resources to the low-speed UE for data transmission efficiency.

[0048] FIG. 4 is a diagram for explaining the operation of a base station according to one embodiment of the present disclosure.

[0049] Referring to FIG. 4, the base station (400) may include a storage unit (410), an inference unit (420), a scheduler (430), and a modem (440).

[0050] The storage unit (410) may include a memory device that stores parameters to be applied to one or more AI-based inference devices included in the inference unit (420). According to one embodiment, the storage unit (410) may store one or more sets of parameters to be applied to one or more AI-based inference devices. According to one embodiment, the storage unit (410) may transmit a set of parameters corresponding to the input of a control signal to the inference unit (420).

[0051] The inference unit (420) may include one or more AI-based inference devices. In one embodiment, the AI-based inference devices may be implemented as AI-based UE speed classifiers. Each AI-based inference device may have the same structure or different structures. The parameter values ​​applied to the AI-based inference devices may be fixed and used, or parameters stored in the storage unit (410) may be used. In one embodiment, existing parameters may be updated with new parameters acquired from the storage unit (410) during the operation of the base station (400) according to conditions. The parameters applied to each AI-based inference device may have the same value or different values.

[0052] According to one embodiment, the inference unit (420) may perform preprocessing on an input signal including at least one of PMI, CQI, and RI. According to another embodiment, a signal input to the inference unit (420) may undergo a preprocessing process before being input to the inference unit (420), and the preprocessed signal may become an input to the inference device.

[0053] At least one AI-based inference device constituting the inference unit (420) can infer (or estimate) whether the UE's speed is high or low based on a preprocessed signal. The inference unit (420) can synthesize the inference results of the at least one AI-based inference device to determine the UE's speed characteristics and transmit the determined UE speed characteristics to the scheduler (430).

[0054] The scheduler (430) may allocate resources to UEs by considering the UE speed characteristics received from the inference unit (420). In one embodiment, the scheduler (430) may allocate resources according to DMRS type 1 to high-speed UEs for accurate channel estimation. In one embodiment, the scheduler (430) may allocate resources according to DMRS type 2 to low-speed UEs for data transmission efficiency.

[0055] The modem (440) can transmit a signal to the UE based on the resource allocation determined by the scheduler (430). The modem (440) can refer to a device that modulates and transmits a signal for information transmission (mainly digital information) and demodulates the signal to restore it to the original signal at the receiving end.

[0056] FIG. 5 illustrates an example of a transmission and reception procedure between a base station and a UE for determining a speed of the UE according to one embodiment of the present disclosure.

[0057] Referring to FIG. 5, the base station (510) can receive at least one of PMI, CQI, and RI from the UE (520) in the (n-L+1)th slot. The UE (520) can perform data decoding in the (n-1)th slot and transmit at least one of PMI, CQI, and RI to the base station (510) in the (n)th slot.

[0058] The base station (510) can classify the speed of the UE (520) based on at least one of the PMI, CQI, and RI received in the (n)th slot. According to one embodiment, the base station (510) can determine the UE (520) as a high-speed UE if the speed of the UE (520) inferred based on at least one of the PMI, CQI, and RI received in the (n)th slot is greater than (or equal to) a threshold value. According to one embodiment, the base station (510) can determine the UE (520) as a low-speed UE if the speed of the UE (520) inferred based on at least one of the PMI, CQI, and RI received in the (n)th slot is less than (or equal to) a threshold value.

[0059] The base station (510) can perform resource allocation for a signal to be transmitted to the UE (520) and / or a signal to be received from the UE (520) based on at least one of the PMI, CQI, and RI received in the (n)th slot and the speed classification of the UE (520).

[0060] The UE (520) can perform data decoding in the (n+k)th slot and transmit at least one of PMI, CQI, and RI to the base station (510) in the (n+k)th slot.

[0061] FIG. 6 is a diagram for explaining the operation of an AI-based UE speed classifier according to one embodiment of the present disclosure.

[0062] In 3GPP standard TS 38.214, CSI-RS codebook can be divided into Type I codebook and Type II codebook. When using Type I codebook, PMI has four elements (i 11 ,i 12 ,i 13 ,i2) and when using Type II codebook, PMI components can be implemented in a variety of ways compared to Type I.

[0063] Figure 6 shows the input and output for an AI-based UE speed classifier when using a Type I codebook. In this disclosure, i 11 ,i 12 ,i 13 ,i2 each PMI 11 , PMI 12 , PMI 13 , expressed as PMI2, and the PMI, CQI, and RI values ​​included in the nth CSI Report are expressed as PMI(n), CQI(n), RI(n), respectively. In Fig. 6, a case for Type I codebook is illustrated as a simple example for convenience of explanation, but the present disclosure is also applicable to Type II codebook. For example, the PMI components of Type II codebook are PMI1, PMI2, ..., PMI K When this is the case, the input of the AI-based UE speed classifier can be expressed as PMI1(n),...,PMIK(n), CQI(n), RI(n).

[0064] In FIG. 6, the base station can determine the speed characteristics of the current UE as high or low using an AI-based inference device (or AI-based UE speed classifier) ​​that takes as input the PMI, CQI, and RI reported as L CSI Reports. According to one embodiment, the AI-based inference device (or AI-based UE speed classifier) ​​can be implemented with a deep neural network structure (or various AI models) including at least one of a Deep Neural Network (DNN) layer, a Recurrent Neural Network (RNN) layer, a Convolution Neural Network (CNN) layer, a Fully Connected (FC) neural network layer, and a Transformer layer.

[0065] Referring to FIGS. 4 and 6, the base station (400) can store PMI, CQI, and RI values ​​received for each CSI report period in the storage unit (410). When L pairs of PMI, CQI, and RI are collected in the storage unit (410), the base station (400) can transmit L pairs of PMI, CQI, and RI to the inference unit (420 or 620). For example, L pairs of PMI, CQI, and RI can include PMI, CQI, and RI values ​​from the (n-L+1)th CSI Report to the (n)th CSI Report. According to one embodiment, L can also be referred to as a window length.

[0066] In one embodiment, L pairs of PMI, CQI, and RI can be expressed in the form of a matrix. For example, the (n-L+1)th CSI Report is PMI 11 (n-L+1), PMI 12 (n-L+1), PMI 13 (n-L+1), PMI2(n-L+1), CQI(n-L+1), RI(n-L+1). The (n-1)th CSI Report may include PMI 11 (n-1), PMI 12 (n-1), PMI 13 (n-1), PMI2(n-1), CQI(n-1), RI(n-1). The (n)th CSI Report may include PMI 11 (n), PMI 12 (n), PMI 13 (n), PMI2(n), CQI(n), and RI(n).

[0067] The inference unit (420 or 620) may perform preprocessing on the input PMI, CQI, and RI. In one embodiment, the preprocessing may include a normalization process that sets the ranges of the PMI, CQI, and RI values ​​to be the same; and a signal distortion process that arbitrarily modifies the input data values.

[0068] In one embodiment, the normalization process can be implemented in various ways. For example, as an example, normalization for the nth PMI, CQI, and RI is as shown in Mathematical Formula 1 below.

[0069] [Mathematical Formula 1]

[0070]

[0071] Based on the above mathematical expression 1, the nth PMI, CQI, and RI input values ​​can all be expressed within the range of 0 to 1.

[0072] The inference unit (420 or 620) can infer whether the UE is high-speed or low-speed based on the input PMI, CQI, and RI.

[0073] FIG. 7 illustrates an example of a base station including multiple AI-based UE speed classifiers according to one embodiment of the present disclosure.

[0074] Referring to FIG. 7, a base station (700) may include a normalization unit (710), a signal distortion unit (720), a plurality of AI-based UE speed classifiers (730, 740, 750), and a determination of UE speed class unit (760).

[0075] The normalization unit (710) can set the range of multiple input data (e.g., at least one of the PMI, CQI, and RI input values) to be the same. For example, the normalization unit (710) can scale all PMI, CQI, and RI values ​​to values ​​between 0 and 1.

[0076] The signal distortion unit (720) can control the data to have different values ​​while maintaining the tendency of the input data by setting small changes to the original input data. In one embodiment, the signal distortion unit (720) can set the data to have different values ​​by adding an arbitrary distortion signal to the original input signal.

[0077] For example, to maintain the trend of the original signal, the PMI of the original signal 11 If these 2 are present, the PMI of the distorted signal 11 can be a value close to 2, such as 2.3. For example, to maintain the trend of the original signal, the PMI of the original signal 11 If these 2 are present, the PMI of the distorted signal 11 may not be distorted to a signal close to 3, such as 2.8.

[0078] According to one embodiment, the signal distortion unit (720) checks the maximum value (MAX) for each of the PMI, CQI, and RI values, and the range of values ​​of the distortion signal (X) is set so as not to damage the tendency of the normalized input signal. can be set to .

[0079] According to one embodiment, the signal distortion unit (720) uses a uniform distribution method. A distortion signal can be generated with any random value within the input. At this time, the probability of occurrence of each value for the distortion signal is 1 / MAX.

[0080] According to one embodiment, the signal distortion unit (720) can generate a distortion signal using a normal distribution method. Due to the characteristics of the normal distribution, the mean is 0 and the variance is When , the probability that the generated distorted signal value exists within the range of -3σ≤X≤3σ can be 99.7%. Therefore, If you set it to , there is a 99.7% chance Values ​​within the range are generated and signal distortion is possible without damaging the tendency of the input signal.

[0081] Each of the plurality of AI-based UE speed classifiers (730, 740, 750) can independently infer the speed of the UE based on input data (for example, at least one of PMI, CQI, and RI input values). The plurality of AI-based UE speed classifiers (730, 740, 750) can include N independent AI-based UE speed classifiers. Each of the plurality of AI-based UE speed classifiers (730, 740, 750) can have the same structure or different structures. The same parameter (or parameter set) can be applied to each of the plurality of AI-based UE speed classifiers (730, 740, 750), or different parameters (or parameter sets) can be applied to each of the plurality of AI-based UE speed classifiers.

[0082] The first AI-based UE speed classification unit (730) receives a normalized input signal (IN0) through the normalization unit (710), and can determine whether the speed of the UE is low (Low) or high (High) based on the normalized input signal (IN0).

[0083] The second AI-based UE speed classification unit (740) receives an input signal (IN1) generated through a normalization unit (710) and a signal distortion unit (720), and can determine whether the speed of the UE is low (Low) or high (High) based on the generated input signal (IN1).

[0084] The N AI-based UE speed classification unit (750) generates an input signal (IN) through a normalization unit (710) and a signal distortion unit (720). N-1 ) and receives the generated input signal (IN N-1 ) can be used to determine whether the UE's speed is low or high.

[0085] The UE speed class determination unit (760) can finally determine whether the speed of the UE is low or high based on the judgment results of each of the multiple AI-based UE speed classification units (730, 740, 750).

[0086] According to one embodiment, if any one of the plurality of AI-based UE speed classifiers (730, 740, 750) determines the speed of the UE as low (Low), the UE speed class determination unit (760) may finally determine the speed of the UE as low (Low). According to one embodiment, if more than half of the plurality of AI-based UE speed classifiers (730, 740, 750) determine the speed of the UE as low (Low), the UE speed class determination unit (760) may finally determine the speed of the UE as low (Low). According to one embodiment, if a set number or more of the plurality of AI-based UE speed classifiers (730, 740, 750) determine the speed of the UE as low (Low), the UE speed class determination unit (760) may finally determine the speed of the UE as low (Low).

[0087] According to one embodiment, if any one of the plurality of AI-based UE speed classifiers (730, 740, 750) determines the speed of the UE as high (High), the UE speed class determination unit (760) may finally determine the speed of the UE as high (High). According to one embodiment, if more than half of the plurality of AI-based UE speed classifiers (730, 740, 750) determine the speed of the UE as high (High), the UE speed class determination unit (760) may finally determine the speed of the UE as high (High). According to one embodiment, if a set number or more of the plurality of AI-based UE speed classifiers (730, 740, 750) determine the speed of the UE as high (High), the UE speed class determination unit (760) may finally determine the speed of the UE as high (High).

[0088] FIG. 8 illustrates an example of a distorted signal having a normal distribution according to one embodiment of the present disclosure.

[0089] Referring to FIGS. 7 and 8, the standard deviation of the distortion signal generated by the signal distortion unit (720) is When set to , the probability that the generated distortion signal value exists within the range of -3σ≤X≤3σ can be 99.7%. Therefore, If you set it to , there is a 99.7% chance Values ​​within the range are generated and signal distortion is possible without damaging the tendency of the input signal.

[0090] FIG. 9 illustrates an example of a change in an input signal after signal distortion according to one embodiment of the present disclosure.

[0091] Fig. 9 shows the input PMI after signal distortion. 11 is an example, and the distorted signal can be controlled so that it does not exceed the normalized range. Referring to Fig. 9, the original PMI 11A signal (IN0) can be distorted to generate a first distorted signal (IN1) and a second distorted signal (IN2). Mathematical expression 2 shows the input PMI after signal distortion when the input range is set between MIN and MAX through normalization. 11 Here is an example of how to avoid exceeding this range.

[0092] [Equation 2]

[0093]

[0094] The N-1 signals generated through signal distortion can be input to each inference device (or AI-based UE speed classification unit). According to one embodiment, at least one of the plurality of inference devices constituting the inference device (or AI-based UE speed classification unit) can receive an undistorted original signal as input. According to one embodiment, the order of the normalization process and the signal distortion process for the input signal can be changed. According to one embodiment, if the structures and application parameters of the inference devices are different, the signal distortion process can be omitted.

[0095] In one embodiment, the speed characteristic of the UE may be selected by synthesizing the results of the inference devices. In one embodiment, the number of inference devices that determine the speed of the UE to be high (N HS ) is the threshold (T HS ) exceeds (N HS > T HS ), the base station can quickly determine the speed of the UE.

[0096] According to one embodiment, the base station T HS By setting to N / 2, if more than half of the inference devices judge as fast, the UE can be judged as fast. According to one embodiment, T HS The base station may determine the UE as a fast UE if any of the inference devices determines the UE as a fast UE by setting T to 0. According to one embodiment, HScan be set to values ​​from 0 to N-1.

[0097] According to one embodiment, the training data set may be composed of PMI, CQI, and RI included in the n-th CSI Report from the (n-L+1)-th CSI Report in the same format as the input data. According to one embodiment, the base station may label the data as 1 if the UE speed at the time of the n-th CSI report exceeds a threshold value, and may label the data as 0 if the UE speed at the time of the n-th CSI report does not exceed the threshold value. According to one embodiment, the base station may perform parameter learning to minimize the loss function through various gradient descent methods.

[0098] In one embodiment, the learned model parameters may be stored in a storage unit within the base station and then transmitted to the inference unit based on an input control signal. In one embodiment, the storage unit may store one or more parameter sets. In one embodiment, the parameter sets may vary depending on the CSI-RS codebook type. In one embodiment, the storage unit may select one of the parameter sets stored in the storage unit and transmit it to the inference unit based on an input control signal including the CSI-RS codebook type.

[0099] FIG. 10 illustrates an example of an inference result of an AI model according to one embodiment of the present disclosure.

[0100] In Fig. 10, an example of the results of determining the speed characteristics of a UE by training an AI model with a labeled data set by setting the speed threshold value, which serves as the standard for high / low speed of a UE, to 5 km / h for the PMI, CQI, and RI inputs proposed in the present disclosure is shown.

[0101] The X-axis of Fig. 10 represents the time at which the CSI report is input, and the Y-axis represents the speed of the UE. In Fig. 10, if the UE is determined to be a high-speed UE, it may be displayed at the 5 km / h point, and if it is determined to be a low-speed UE, it may be displayed at the 0 km / h point. At this time, it may be possible to determine with high accuracy whether the UE is high-speed or low-speed from the PMI, CQI, and RI information acquired from the CSI report.

[0102] In one embodiment, the base station may include multiple inference devices (or AI-based UE speed classifiers) having the same structure and parameters. In one embodiment, when the input signal to the inference device (or AI-based UE speed classifier) ​​is on the ambiguous boundary for determining whether the UE is at a high or low speed, the base station may alleviate the ambiguity in inference by randomly generating signals with a similar tendency to the input signal through signal distortion and making a determination for each signal.

[0103] [Table 1]

[0104]

[0105] Table 1 is an example showing the performance according to the number of classifiers (Number of Classifiers) of inference devices (or AI-based UE speed classifiers) proposed in the present disclosure. Detection rate refers to the rate at which UEs are correctly estimated as high / low speeds, and Error rate refers to the rate at which speed characteristics are incorrectly determined as low speeds among high-speed UEs. Referring to Table 1, it can be confirmed that as the number of classifiers (Number of Classifiers) of inference devices (or AI-based UE speed classifiers) increases, performance gains occur in terms of Detection rate and Error rate.

[0106] From an implementation perspective, using multiple inference units (or AI-based UE speed classifiers) increases the inference complexity compared to using only one inference unit (or AI-based UE speed classifier). However, since the inference units operate independently, parallelization is possible, so the increase in latency can be significantly reduced. Furthermore, since all inference units (or AI-based UE speed classifiers) are configured with the same structure and parameters, performance gains can be achieved even with a single parameter set. Therefore, compared to using only one inference unit (or AI-based UE speed classifier), no additional memory is required for parameter storage in the storage, and control operations for the storage can be simplified because the same parameter set can be selected.

[0107] FIG. 11 is a diagram illustrating the operation of a base station according to one embodiment of the present disclosure. FIG. 11 illustrates the operation of a base station including an AI model that infers speed characteristics based on PMI, CQI, and RI information.

[0108] In operation 1101, the base station may receive a CSI report from the UE. In operation 1103, the base station may parse the PMI, CQI, and RI from the CSI report received from the UE. In operation 1105, the base station may store the PMI, CQI, and RI obtained from the CSI report in a storage unit within the base station.

[0109] In operation 1107, the base station can check whether there are L pairs of PMI, CQI, and RI stored in the storage unit. If there are not L pairs of PMI, CQI, and RI stored in the storage unit (1107-NO), the base station can perform operations 1101 to 1105 again. If there are L pairs of PMI, CQI, and RI stored in the storage unit (1107-YES), in operation 1109, the base station can transfer L pairs of PMI, CQI, and RI from the storage unit in the base station to the inference unit in the base station.

[0110] In operation 1111, the base station can classify whether the UE's speed is low or high using the inference unit. In operation 1113, the base station can determine whether the UE is a low-speed UE. If the UE is a low-speed UE (1113-YES), in operation 1115, the base station can set DMRS configuration type 2 for the UE. If the UE is not a low-speed UE (1113-NO), in operation 1117, the base station can set DMRS configuration type 1 for the UE.

[0111] DMRS configuration Type 1 has the advantage of achieving superior channel estimation performance compared to DMRS configuration Type 2 due to dense allocation of DRMS, but has the disadvantage of not being able to transmit as much PDSCH data. On the other hand, DMRS configuration Type 2 can transmit a lot of PDSCH data, but may suffer from channel estimation performance. If the UE is determined to be slow, the base station may configure DMRS configuration Type 2 to gain data transmission rather than channel estimation accuracy, assuming that the channel change will not be abrupt. If the UE is determined to be fast, the base station may configure DMRS configuration Type 1 to improve channel estimation accuracy, assuming that the channel change will be abrupt.

[0112] According to the 3GPP standard, the DMRS configuration type can be determined by the dmrs-Type field value of the RRC parameter DMRS-DownlinkConfig. In one embodiment, if the dmrs-Type field value is set to 'type2', it can be set to DMRS configuration type 2, and if no value is assigned to the field, it can be set to DMRS configuration type 1.

[0113] In one embodiment, the base station may perform frequency diverse scheduling to maximize reception reliability when the UE is determined to be a high-speed UE. In one embodiment, the base station may perform localized frequency scheduling to maximize system spectral efficiency when the UE is determined to be a low-speed UE.

[0114] In one embodiment, a method of a base station may include receiving channel state information (CSI) including a precoding matrix indicator (PMI), a channel quality indicator (CQI), and a rank indicator (RI) from a user equipment (UE). In one embodiment, the method of the base station may include performing a normalization procedure to adjust the values ​​of each of the PMI, the CQI, and the RI to fall within a preset range. In one embodiment, the method of the base station may include determining, through at least one inference unit included in the base station, whether a speed of the UE is high or low using the normalized PMI, CQI, and RI. In one embodiment, the method of the base station may include allocating resources to the UE based on whether the speed of the UE is high or low.

[0115] In one embodiment, the method of the base station may include an operation of generating a distortion signal by adding a random value to at least one value among normalized PMI, CQI, and RI. In one embodiment, the method of the base station may further include an operation of determining, through at least one inference unit included in the base station, whether the speed of the UE is high or low using the distortion signal and the normalized PMI, CQI, and RI.

[0116] According to one embodiment, the inference device may be configured with a deep neural network structure including at least one of a DNN (Deep Neural Network) layer, an RNN (Recurrent Neural Network) layer, a CNN (Convolution Neural Network) layer, an FC (Fully Connected) neural network layer, and a Transformer layer.

[0117] In one embodiment, the method of the base station may include an operation of checking whether L pairs of PMI, CQI, and RI are stored in a storage unit within the base station. In one embodiment, the method of the base station may include an operation of transferring L pairs of PMI, CQI, and RI from the storage unit within the base station to the inference device within the base station, if the L pairs of PMI, CQI, and RI are stored in the storage unit.

[0118] In one embodiment, the method of the base station may include an operation of determining whether the speed of the UE is high or low through a plurality of inference devices included in the base station using normalized PMI, CQI, and RI. In one embodiment, the method of the base station may further include an operation of determining the speed of the UE to be high if a set number or more of the plurality of inference devices determine that the speed of the UE is high.

[0119] In one embodiment, the method of the base station may include an operation of determining whether the speed of the UE is high or low through a plurality of inference devices included in the base station using normalized PMI, CQI, and RI. In one embodiment, the method of the base station may further include an operation of determining the speed of the UE as low if a set number or more of the plurality of inference devices determine that the speed of the UE is low.

[0120] According to one embodiment, the method of the base station may include an operation of allocating DMRS configuration type 1 to the UE if the speed of the UE is determined to be high. According to one embodiment, the method of the base station may further include an operation of allocating DMRS configuration type 2 to the UE if the speed of the UE is determined to be low.

[0121] According to one embodiment, a storage medium storing at least one computer-readable instruction may cause the at least one instruction, when executed by at least a part of at least one processor (120) of an electronic device, to cause the electronic device to perform at least one operation. The at least one operation may include receiving, from a user equipment (UE), channel state information (CSI) including a precoding matrix indicator (PMI), a channel quality indicator (CQI), and a rank indicator (RI). The at least one operation may include performing a normalization procedure to adjust the values ​​of each of the PMI, the CQI, and the RI to within a preset range. The at least one operation may include determining, through at least one inference unit included in the base station, whether a speed of the UE is high or low using the normalized PMI, CQI, and RI. The at least one operation may include allocating resources to the UE based on whether the speed of the UE is high or low.

[0122] According to one embodiment, an electronic device includes at least one processor (120); and a memory (130) storing at least one instruction, wherein the at least one instruction, when executed by at least a part of the at least one processor (120), causes the electronic device to perform at least one operation. The at least one operation may include receiving, from a user equipment (UE), channel state information (CSI) including a precoding matrix indicator (PMI), a channel quality indicator (CQI), and a rank indicator (RI). The at least one operation may include performing a normalization procedure to adjust the values ​​of each of the PMI, the CQI, and the RI to within a preset range. The at least one operation may include determining, through at least one inference unit included in the base station, whether a speed of the UE is high or low by using the normalized PMI, CQI, and RI. The at least one operation may include an operation of allocating resources to the UE depending on whether the UE has a high speed or a low speed.

[0123] The embodiments of this document and the terminology used herein are not intended to limit the technical features described in this document to specific embodiments, but should be understood to include various modifications, equivalents, or substitutes of the 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 of the items, unless the context clearly indicates otherwise. In this document, each of the phrases "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" can include any one of the items listed together in the corresponding phrase among those phrases, or all possible combinations thereof. Terms such as "first," "second," or "first" or "second" may be used merely to distinguish one component from another, and do not limit the components in any other respect (e.g., importance or order). When a component (e.g., a first component) is referred to as "coupled" or "connected" to another (e.g., a second component), with or without the terms "functionally" or "communicatively," it means that the component can be connected to the other component directly (e.g., wired), wirelessly, or through a third component.

[0124] The term "module" used in the embodiments of this document 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 an integral component, or a minimum unit or part of such a component 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).

[0125] One embodiment of the present document may be implemented as software (e.g., a program (140)) including one or more instructions stored in a storage medium (e.g., an internal memory (136) or an external memory (138)) readable by a machine (e.g., an electronic device (101)). For example, a processor (e.g., a processor (120)) of the machine (e.g., an electronic device (101)) may call at least one instruction among 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 the at least one called instruction. The one or more instructions may include code generated by a compiler or code executable by an interpreter. The machine-readable storage medium 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 signals (e.g., electromagnetic waves), and the term does not distinguish between cases where data is stored semi-permanently or temporarily on the storage medium.

[0126] According to one embodiment, the method according to one embodiment disclosed in this document may be provided as included in a computer program product. The computer program product may be traded as a product between a seller and a buyer. The computer program product may be distributed in the form of a machine-readable storage medium (e.g., compact disc read-only memory (CD-ROM)), or may be distributed online (e.g., downloaded or uploaded) via an application store (e.g., Play Store™) or directly between two user devices (e.g., smart phones). In the case of online distribution, at least a portion of the computer program product may be temporarily stored or temporarily generated in a machine-readable storage medium, such as the memory of a manufacturer's server, an application store's server, or an intermediary server.

[0127] According to one embodiment, each component (e.g., a module or a program) of the above-described components may include one or more entities, and some of the entities may be separated and arranged in other components. According to one embodiment, one or more components or operations of the aforementioned components may be omitted, or one or more other components or operations may be added. Alternatively or additionally, a plurality of components (e.g., a module or a program) may be integrated into a single component. In this case, the integrated component may perform one or more functions of each of the plurality of components identically or similarly to those performed by the corresponding component among the plurality of components prior to the integration. According to one embodiment, the operations performed by a module, program, or other component may be executed sequentially, in parallel, iteratively, or heuristically, or one or more of the operations may be executed in a different order, omitted, or one or more other operations may be added.

Claims

In a method of a base station in a wireless communication system, An operation of receiving channel state information (CSI) including a precoding matrix indicator (PMI), a channel quality indicator (CQI), and a rank indicator (RI) from a user equipment (UE); An operation of performing a normalization procedure to adjust the values ​​of each of the PMI, the CQI, and the RI to within a preset range; An operation of determining whether the speed of the UE is high or low using normalized PMI, CQI, and RI through at least one inference unit included in the base station; and A method characterized by including an operation of allocating resources to the UE depending on whether the speed of the UE is high or low. In the first paragraph, An operation of generating a distorted signal by adding a random value to at least one of the normalized PMI, CQI, and RI; and A method characterized in that it further includes an operation of determining whether the speed of the UE is high or low by using the distorted signal and the normalized PMI, CQI, and RI through at least one inference unit included in the base station. In the first paragraph, A method characterized in that the above inference device is configured with a deep neural network structure including at least one of a DNN (Deep Neural Network) layer, an RNN (Recurrent Neural Network) layer, a CNN (Convolution Neural Network) layer, an FC (Fully Connected) neural network layer, and a Transformer layer. In the first paragraph, An operation of checking whether the L pair of PMI, CQI, and RI is stored in the storage unit within the base station; and A method characterized in that, when the L pair of PMI, CQI, and RI is stored in the storage unit, the method further includes an operation of transmitting the L pair of PMI, CQI, and RI from the storage unit in the base station to the inference device in the base station. In the first paragraph, An operation of determining whether the speed of the UE is high or low through a plurality of inference devices included in the base station using normalized PMI, CQI, and RI; and A method characterized in that it further includes an operation of determining that the speed of the UE is high speed when a set number or more of the plurality of inference devices determine that the speed of the UE is high speed. In the first paragraph, An operation of determining whether the speed of the UE is high or low through a plurality of inference devices included in the base station using normalized PMI, CQI, and RI; and A method characterized in that it further includes an operation of determining that the speed of the UE is low when a set number or more of the plurality of inference devices determine that the speed of the UE is low. In the first paragraph, An operation of allocating DMRS configuration type 1 to the UE when the speed of the UE is determined to be high speed; or A method characterized in that it further includes an operation of allocating DMRS configuration type 2 to the UE when the speed of the UE is determined to be low. A storage medium storing at least one computer-readable instruction, wherein the at least one instruction, when executed by at least a part of at least one processor of a base station, causes the base station to perform at least one operation; At least one of the above actions: An operation of receiving channel state information (CSI) including a precoding matrix indicator (PMI), a channel quality indicator (CQI), and a rank indicator (RI) from a user equipment (UE); An operation of performing a normalization procedure to adjust the values ​​of each of the PMI, the CQI, and the RI to within a preset range; An operation of determining whether the speed of the UE is high or low using normalized PMI, CQI, and RI through at least one inference unit included in the base station; and A storage medium characterized by including an operation of allocating resources to the UE depending on whether the speed of the UE is high or low. In the 8th paragraph, the at least one operation: An operation of generating a distorted signal by adding a random value to at least one of the normalized PMI, CQI, and RI; and A storage medium characterized in that it further includes an operation of determining whether the speed of the UE is high or low by using the distorted signal and the normalized PMI, CQI, and RI through at least one inference unit included in the base station. In paragraph 8, A storage medium characterized in that the above inference device is configured with a deep neural network structure including at least one of a DNN (Deep Neural Network) layer, an RNN (Recurrent Neural Network) layer, a CNN (Convolution Neural Network) layer, an FC (Fully Connected) neural network layer, and a Transformer layer. In the 8th paragraph, the at least one operation: An operation of checking whether the L pair of PMI, CQI, and RI is stored in the storage unit within the base station; and A storage medium characterized in that, when the L pair of PMI, CQI, and RI is stored in the storage unit, the storage medium further includes an operation of transmitting the L pair of PMI, CQI, and RI from the storage unit in the base station to the inference device in the base station. In the 8th paragraph, the at least one operation: An operation of determining whether the speed of the UE is high or low through a plurality of inference devices included in the base station using normalized PMI, CQI, and RI; and A storage medium characterized in that it further includes an operation of determining that the speed of the UE is high speed when a set number or more of the plurality of inference devices determine that the speed of the UE is high speed. In the 8th paragraph, the at least one operation: An operation of determining whether the speed of the UE is high or low through a plurality of inference devices included in the base station using normalized PMI, CQI, and RI; and A storage medium characterized in that it further includes an operation of determining the speed of the UE as low when a set number or more of the plurality of inference devices determine that the speed of the UE is low. At the base station, At least one processor (120); and Contains a memory (130) storing at least one instruction, wherein said at least one instruction, when executed by at least a portion of said at least one processor, causes said base station to perform at least one operation; At least one of the above actions: An operation of receiving channel state information (CSI) including a precoding matrix indicator (PMI), a channel quality indicator (CQI), and a rank indicator (RI) from a user equipment (UE); An operation of performing a normalization procedure to adjust the values ​​of each of the PMI, the CQI, and the RI to within a preset range; An operation of determining whether the speed of the UE is high or low using normalized PMI, CQI, and RI through at least one inference unit included in the base station; and A base station characterized by including an operation of allocating resources to the UE depending on whether the speed of the UE is high or low. In paragraph 14, the at least one operation: An operation of generating a distorted signal by adding a random value to at least one of the normalized PMI, CQI, and RI; and A base station characterized in that it further includes an operation of determining whether the speed of the UE is high or low through at least one inference unit included in the base station using the distorted signal and the normalized PMI, CQI, and RI.

Images