Beam management method and device using artificial intelligence and / or machine learning

Abstract

A method and an apparatus for performing beam management in a wireless communication system are provided. A terminal triggers at least one beam failure recovery (BFR) for a cell that performs beam management using an artificial intelligence and / or machine learning model, and performs inactivation of the artificial intelligence and / or machine learning model based on the number of times of the triggered at least one beam failure recovery within a specific time period. And then, inactivation information of the artificial intelligence and / or machine learning model is transmitted to the base station. And then, inactivation information of the artificial intelligence and / or machine learning model is transmitted to the base station. Meanwhile, the base station transmits setting information related to the artificial intelligence and / or the machine learning model to the terminal, and receives inactivation information of the artificial intelligence and / or the machine learning model from the terminal. Beam generation related to artificial intelligence and / or machine learning models is then interrupted based on the received inactivation information.

Description

Technical Field

[0001] This specification relates to the 3GPP 5G NR system. Background Art

[0002] As the times change, more and more communication devices require higher throughput, demanding wireless broadband communications that surpass existing LTE systems—the next-generation 5G system. Within this next-generation 5G system, known as NewRAT, communication scenarios are categorized into enhanced mobile broadband (eMBB), ultra-reliability and low-latency communication (URLLC), and massive machine-type communications (mMTC).

[0003] Among them, eMBB is the next-generation mobile communication scenario with the characteristics of high spectrum efficiency, high user experience data rate, high peak data rate, etc.; URLLC is the next-generation mobile communication scenario with the characteristics of ultra-reliability, ultra-low latency, ultra-high availability, etc. (for example, V2X, emergency services, remote control); mMTC is the next-generation mobile communication scenario with the characteristics of low cost, low energy consumption, short data packets, and massive connections (for example, IoT). Summary of the Invention

[0004] Technical issues

[0005] An object of the present disclosure is to provide a method and apparatus for effectively managing an AI / ML model in response to beam failure recovery (BFR) in a wireless communication system when the model is applied to a beam management process.

[0006] Technical Solution

[0007] One embodiment of the present specification provides a method, in which, in a wireless communication system, a terminal triggers at least one beam failure recovery (BFR) for a cell that performs beam management using an artificial intelligence and / or machine learning model, deactivates the artificial intelligence and / or machine learning model based on the number of times the at least one beam failure recovery is triggered within a specific time interval, and then sends deactivation information of the artificial intelligence and / or machine learning model to a base station.

[0008] In addition, one embodiment of the present specification provides a method, in which, in a wireless communication system, a base station sends setting information related to an artificial intelligence and / or machine learning model to a terminal, receives deactivation information of the artificial intelligence and / or machine learning model from the terminal, and then, based on the received deactivation information, interrupts beamforming related to the artificial intelligence and / or machine learning model.

[0009] In addition, an embodiment of the present invention provides a communication device, which includes: at least one processor in a wireless communication system; and at least one memory, storing instructions and operably electrically connected to the at least one processor, and executed by the at least one processor based on the instructions, and the operations performed are: for a cell that performs beam management (beam failure recovery, BFR) using artificial intelligence and / or machine learning models, triggering at least one beam failure recovery, deactivating the artificial intelligence and / or machine learning model based on the number of times at least one beam failure recovery is triggered within a specific time interval, and then sending deactivation information of the artificial intelligence and / or machine learning model to a base station.

[0010] In addition, an embodiment of the present invention provides a base station, which includes: at least one processor in a wireless communication system; and at least one memory, which stores instructions and is operably electrically connected to the at least one processor, and is executed by the at least one processor based on the instructions, and the operations performed are: sending setting information related to artificial intelligence and / or machine learning models to a terminal, receiving deactivation information of the artificial intelligence and / or machine learning models from the terminal, and then, based on the received deactivation information, interrupting beamforming related to the artificial intelligence and / or machine learning models.

[0011] The terminal may start a timer associated with the deactivation of the artificial intelligence and / or machine learning model, and during the driving period of the started timer, the number of times the triggered at least one beam failure is recovered may be counted.

[0012] In addition, when the counted number of at least one beam failure recovery reaches a maximum number, the terminal may stop the timer. In addition, when the timer expires, the counted number of at least one beam failure recovery may be initialized.

[0013] On the one hand, the terminal can receive setting information related to the artificial intelligence and / or machine learning model from the base station, but the setting information related to the artificial intelligence and / or machine learning model may include at least one of timer value information and maximum number of beam failure recovery information.

[0014] Furthermore, the deactivation information sent by the terminal to the base station may include at least one of the following: i) identity (ID) information of the artificial intelligence and / or machine learning model; ii) information notifying that the artificial intelligence and / or machine learning model is deactivated; and iii) an instruction instructing the deactivation of the artificial intelligence and / or machine learning model of the base station.

[0015] On the one hand, based on the deactivation information received from the terminal, the base station can fallback to beam management that does not apply the artificial intelligence and / or machine learning model.

[0016] Effects of the Invention

[0017] According to the disclosure of this specification, when using AI / ML models to perform beam management technology, system performance degradation caused by model performance degradation can be prevented. Specifically, the model is deactivated based on the occurrence / triggering of continuous beam failure recovery (BFR) of the terminal, thereby preventing continuous beam failure caused by the use of unreliable AI / ML models, thereby minimizing the overall system performance degradation. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A diagram illustrating a wireless communication system.

[0019] Figure 2 The structure of the radio frame used by NR is shown.

[0020] Figures 3a to 3c An exemplary diagram illustrating an exemplary architecture for wireless communication services.

[0021] Figure 4 Shows the time slot structure of the NR frame.

[0022] Figure 5 Examples of subframe types in NR are shown.

[0023] Figure 6 Shows the structure of a self-contained time slot.

[0024] Figure 7 Figure 2 shows an example diagram of beam failure detection (BFD) operation in NR.

[0025] Figures 8a to 8b Beam measurement and spatial domain beam prediction using AI / ML are shown.

[0026] Figure 9 Shown is temporal domain beam prediction using AI / ML.

[0027] Figure 10 A method for operating a terminal according to an embodiment of this specification is shown.

[0028] Figures 11a to 11b An example of model deactivation based on model execution according to an embodiment of this specification is shown.

[0029] Figures 12a to 12c An example of model deactivation performed based on a serving cell according to an embodiment of this specification is shown.

[0030] Figure 13 An example of model deactivation based on terminal execution according to an embodiment of this specification is shown.

[0031] Figure 14 An example of operations of a base station and a terminal according to an embodiment of this specification is shown.

[0032] Figure 15 A method for operating a terminal according to an embodiment of this specification is shown.

[0033] Figure 16 The following describes an operating method of a base station according to an embodiment of the present specification.

[0034] Figure 17 A device according to an embodiment of the present specification is shown.

[0035] Figure 18 It is a block diagram showing the structure of a terminal according to an embodiment of this specification.

[0036] Figure 19 A block diagram showing the configuration of a processor that implements the disclosure of this specification.

[0037] Figure 20 To show in detail Figure 17 The transceiver of the first device or Figure 18 Block diagram of the transceiver portion of the device shown. DETAILED DESCRIPTION

[0038] It should be noted that the technical terms used in this specification are only used to describe specific embodiments and are not intended to limit the content of this specification. In addition, as long as they are not specifically defined as different meanings in the present invention, the technical terms used in this specification should be interpreted as the meanings commonly understood by ordinary technicians in the field to which this specification discloses, and shall not be interpreted as overly broad meanings or overly narrow meanings. In addition, when the technical terms used in the specification are inappropriate technical terms that cannot accurately express the content and ideas of this specification, they must be replaced with technical terms that can be correctly understood by technicians. In addition, the common terms used in this specification must be interpreted according to the pre-defined content or context and shall not be interpreted as overly narrow meanings.

[0039] In addition, unless the context clearly indicates otherwise, singular expressions used in this specification include plural expressions. In this application, terms such as "consisting of" or "having" should not be interpreted as necessarily including all of the multiple components or steps described in the specification, but should be interpreted as not including some of the components or steps, or including additional components or steps.

[0040] Furthermore, terms including ordinal numbers such as "1" and "2" used in this specification may be used to describe various components, but these components are not limited by these terms. These terms are used solely to distinguish one component from another. For example, without exceeding the scope of the rights, the first component may be named the second component, and similarly, the second component may be named the first component.

[0041] When a component is referred to as being “connected” or “connected to” another component, it can be directly connected or connected to the other component or with other components intervening. Conversely, when a component is referred to as being “directly connected” or “directly connected to” another component, it should be understood that there are no other components intervening.

[0042] The following describes the embodiments in detail with reference to the accompanying drawings. Regardless of the figure numbers, the same or similar components are given the same figure numbers and their repeated descriptions are omitted. In addition, in describing the content of the present invention, when it is judged that the specific description of the relevant known technology may obscure the gist of this specification, the detailed description will be omitted. In addition, it should be noted that the drawings are only used to make it easier to understand the content and ideas of this specification and should not be interpreted as limiting the content and ideas of this specification by the drawings. The content and ideas of this specification should be interpreted as extending to all changes, equivalents and substitutes in addition to the drawings.

[0043] In this specification, "A or B" may mean "only A," "only B," or "all A and B." In other words, in this specification, "A or B" may be interpreted as "A and / or B." For example, in this specification, "A, B, or C" may mean "only A," "only B," "only C," or "any combination of A, B, and C."

[0044] As used herein, a slash ( / ) or a comma (comma) may mean "and / or." For example, "A / B" may mean "A and / or B." Thus, "A / B" may mean "only A," "only B," or "both A and B." For example, "A, B, C" may mean "A, B, or C."

[0045] In the present specification, “atleast one of A and B” may mean “only A”, “only B” or “all of A and B”. In addition, in the present specification, the expression “at least one of A or B” or “at least one of A and / or B” may be interpreted the same as “at least one of A and B”.

[0046] In addition, in this specification, “atleast one of A, B and C” may mean “only A”, “only B”, “only C”, or “any combination of A, B and C”. In addition, “atleast one of A, B or C” or “atleast oneof A, B and / or C” may mean “at least one of A, B and C”.

[0047] In addition, the brackets used in this specification may mean "for example". Specifically, when identified as "control information (PDCCH)", "PDCCH (physical downlink control channel)" may be proposed as an example of "control information". In other words, the "control information" in this specification is not limited to "PDCCH", and "PDDCH" may be proposed as an example of "control information". In addition, when identified as "control information (i.e., PDCCH)", "PDCCH" may be proposed as an example of "control information".

[0048] In this specification, technical features described individually in one drawing can be implemented independently or simultaneously.

[0049] The accompanying drawings illustrate UE (User Equipment), but the UE shown here may also be replaced by a term such as Terminal or Mobile Equipment. Furthermore, the UE may be a portable device such as a laptop computer, mobile phone, PDA (Personal Digital Assistant), smartphone, multimedia device, or a non-portable device such as a PC or vehicle-mounted device.

[0050] Below, a UE is used as an example of a device capable of wireless communication (e.g., a wireless communication device, a wireless device, or a wireless appliance). The operations performed by a UE can be performed by any device capable of wireless communication. It may also be referred to as a device capable of wireless communication, a wireless communication device, a wireless device, or a wireless appliance.

[0051] The term base station used below generally refers to a fixed station that communicates with wireless devices, and can be used as a broad term including eNodeB (evolved-NodeB), eNB (evolved-NodeB), BTS (Base Transceiver System), access point (Access Point), gNB (Next generation NodeB), RRH (remote radiohead), TP (transmission point), RP (reception point), relay, etc.

[0052] This specification uses the LTE system, LTE-A system, and NR system to describe the embodiments, but these embodiments can also be applied to any communication system to which the definitions apply.

[0053] <Wireless Communication System>

[0054] Benefiting from the success of LTE (long term evolution) / LTE-Advanced (LTE-A) used for fourth-generation mobile communications, the fifth-generation (so-called 5G) mobile communications, which is the next generation, has been commercialized and subsequent research is underway.

[0055] The International Telecommunication Union (ITU) defines fifth-generation mobile communications as providing a maximum data transmission speed of 20Gbps and a minimum perceptible transmission speed of 100Mbps or higher at any location. Its official name is "IMT-2020."

[0056] Three major usage scenarios were proposed in the ITU, such as eMBB (enhanced Mobile BroadBand), mMTC (massive Machine Type Communication) and URLLC (Ultra Reliable and Low Latency Communications).

[0057] URLLC addresses application scenarios requiring high reliability and low latency. For example, autonomous driving, factory automation, and augmented reality services require both high reliability and low latency (e.g., sub-1ms). Current 4G (LTE) latency is statistically 21-43ms (best 10%) and 33-75ms (median). This is insufficient to support services requiring sub-1ms latency. eMBB, on the other hand, addresses mobile ultra-broadband requirements.

[0058] Specifically, the fifth-generation mobile communication system supports higher capacity than the current 4G LTE, increasing the density of mobile broadband users and supporting D2D (Device-to-Device), high stability, and MTC (Machine-type Communication). To better realize the Internet of Things, 5G research and development also aims to achieve lower standby times and lower power consumption than 4G mobile communication systems. To support this 5G mobile communication, new radio access technology (New RAT or NR) can be proposed.

[0059] NR frequency bands can be defined as two types of frequency ranges (FR1 and FR2). The values ​​of the frequency ranges can be changed. For example, the two types of frequency ranges (FR1 and FR2) can be shown in Table 1 below. For ease of description, FR1 in the frequency range used by NR systems can refer to the "sub 6 GHz range" and FR2 can refer to the "above 6 GHz range", which can be referred to as millimeter waves (mmW).

[0060]

Table 1

[0061] Frequency range identification Corresponding frequency range Subcarrier spacing FR1 410MHz-7125MHz 15, 30, 60kHz FR2 24250MHz-52600MHz 60, 120, 240kHz

[0062] The frequency range values ​​of the NR system can be changed. For example, FR1, as shown in Table 1, can include frequency bands from 410 MHz to 7125 MHz. That is, FR1 can include frequency bands above 6 GHz (or 5850, 5900, 5925 MHz, etc.). For example, the frequency bands above 6 GHz (or 5850, 5900, 5925 MHz, etc.) included in FR1 can include unlicensed bands. Unlicensed bands can be used for various purposes, for example, they can be used for vehicle communications (e.g., autonomous driving).

[0063] On the other hand, 3GPP-based communication standards define downlink physical channels corresponding to resource elements that carry information from upper layers, and downlink physical signals corresponding to resource elements used by the physical layer but not carrying information from upper layers. For example, the physical downlink shared channel (PDSCH), physical broadcast channel (PBCH), physical multicast channel (PMCH), physical control format indicator channel (PCFICH), physical downlink control channel (PDCCH), and physical hybrid ARQ indicator channel (PHICH) are defined as downlink physical channels, and reference signals and synchronization signals are defined as downlink physical signals. Reference signals (RS), also known as pilots, are predefined signals with specific waveforms known to both the gNB and the UE. For example, cell-specific RS, UE-specific RS, positioning RS (PRS), and channel state information RS (CSI-RS) are defined as downlink reference signals. The 3GPP LTE / LTE-A standards define uplink physical channels corresponding to resource elements that carry information from higher layers, as well as uplink physical signals corresponding to resource elements used by the physical layer but not carrying information from higher layers. For example, the physical uplink shared channel (PUSCH), physical uplink control channel (PUCCH), and physical random access channel (PRACH) are defined as uplink physical channels. The demodulation reference signal (DMRS) for uplink control / data signals and the sounding reference signal (SRS) for uplink channel measurement are defined.

[0064] In this specification, PDCCH (Physical Downlink Control CHannel) / PCFICH (Physical Control Format Indicator CHannel) / PHICH (Physical Hybrid automatic retransmit request Indicator CHannel) / PDSCH (Physical Downlink Shared CHannel) refer to a set of time-frequency resources or a set of resource elements that carry DCI (downlink control information) / CFI (Control Format Indicator) / downlink ACK / NACK (ACK nowlegement / Negative ACK) / downlink data, respectively. In addition, PUCCH (Physical Uplink Control CHannel) / PUSCH (Physical Uplink Shared CHannel) / PRACH (Physical Random Access CHannel) refer to a set of time-frequency resources or a set of resource elements that carry UCI (Uplink Control Information) / uplink data / random access signals, respectively.

[0065] Figure 1 A diagram illustrating a wireless communication system.

[0066] Reference Figure 1 As can be seen, the wireless communication system includes at least one base station (BS). The BS is divided into gNodeB (or gNB) 20a and eNodeB (or eNB) 20b. The gNB 20a supports fifth-generation mobile communications. The eNB 20b supports fourth-generation mobile communications, namely LTE (long term evolution).

[0067] Each base station 20a and 20b provides communication services for a specific geographical area (generally referred to as a cell) 20-1, 20-2, 20-3. A cell can be further divided into a plurality of areas (referred to as sectors).

[0068] A UE (User Equipment) typically belongs to a cell, which is called a serving cell. The base station that provides communication services to the serving cell is called a serving base station (serving BS). Wireless communication systems are cellular systems, so there are other cells adjacent to the serving cell. These other cells adjacent to the serving cell are called neighboring cells. The base station that provides communication services to a neighboring cell is called a neighboring base station (neighboring BS). The serving cell and neighboring cells are determined relative to the UE.

[0069] Hereinafter, downlink refers to communication from base station 20 to UE 10, and uplink refers to communication from UE 10 to base station 20. In the downlink, the transmitter may be part of base station 20, and the receiver may be part of UE 10. In the uplink, the transmitter may be part of UE 10, and the receiver may be part of base station 20.

[0070] On the other hand, wireless communication systems can be roughly divided into FDD (frequency division duplex) and TDD (time division duplex) modes. According to the FDD mode, uplink transmission and downlink transmission occupy different frequency bands. According to the TDD mode, uplink transmission and downlink transmission occupy the same frequency band and are performed at different times. The channel response of the TDD mode is essentially reciprocal. In a given frequency region, the downlink channel response and the uplink channel response are almost the same. Therefore, in a wireless communication system based on TDD, the downlink channel response has the advantages that can be obtained from the uplink channel response. In the TDD mode, uplink transmission and downlink transmission are time-divided over the entire frequency band, so the downlink transmission of the base station and the uplink transmission of the UE cannot be performed at the same time. In a TDD system in which uplink transmission and downlink transmission are divided into subframe units, uplink transmission and downlink transmission are performed in different subframes.

[0071] Figure 2 The structure of the radio frame used by NR is shown.

[0072] In NR, uplink and downlink transmissions consist of frames. A radio frame has a length of 10ms and is defined as two 5ms half-frames (HF). A half-frame is defined as five 1ms subframes (SF). A subframe is divided into one or more time slots, and the number of time slots in a subframe depends on the SCS (Subcarrier Spacing). Each time slot includes 12 or 14 OFDM (A) symbols depending on the cyclic prefix (CP). When using a normal CP, each time slot includes 14 symbols. When using an extended CP, each time slot includes 12 symbols. Among them, the symbols may include OFDM symbols (or CP-OFDM symbols) and SC-FDMA symbols (or DFT-s-OFDM symbols).

[0073] <Support for various parameter sets (numerology)>

[0074] In the NR system, as wireless communication technology develops, multiple numerologies can also be provided to terminals. For example, when the SCS is 15kHz, it supports wide areas in the traditional cellular band. When the SCS is 30kHz / 60kHz, it supports dense urban areas, lower latency, and wider carrier bandwidth. When the SCS is 60kHz or higher, it supports bandwidths larger than 24.25GHz to overcome phase noise.

[0075] The parameter set can be defined based on the CP (cycle prefix) length and subcarrier spacing (SCS). A cell can provide multiple parameter sets to the terminal. When μ is used to represent the parameter set index, the CP length corresponding to each subcarrier spacing can be as shown in the following table.

[0076]

Table 2

[0077] μ <![CDATA[Δf=2 μ ·15[kHz]]]> CP 0 15 ordinary 1 30 ordinary 2 60 Normal, Extended 3 120 ordinary 4 240 ordinary 5 480 ordinary 6 960 ordinary

[0078] For a common CP, when μ represents the index of the parameter set, the number of OFDM symbols per time slot (N slot symb ), the number of time slots per frame (N frame,μ slot ) and the number of time slots in each subframe (N subframe,μ slot ) are shown in the following table.

[0079]

Table 3

[0080] μ <![CDATA[Δf=2 μ ·15[kHz]]]> <![CDATA[N slot symb ]]> <![CDATA[N frame，μ slot ]]> <![CDATA[N subframe,μ slot ]]> 0 15 14 10 1 1 30 14 20 2 2 60 14 40 4 3 120 14 80 8 4 240 14 160 16 5 480 14 320 32 6 960 14 640 64

[0081] For extended CP, when μ represents the index of the parameter set, the number of OFDM symbols per time slot (N slot symb ), the number of time slots per frame (N frame,μ slot ) and the number of time slots in each subframe (N subframe,μ slot ) are shown in the following table.

[0082]

Table 4

[0083]

[0084] In NR systems, OFDM(A) parameter sets (e.g., SCS, CP length, etc.) can be configured differently between multiple cells combined into a single terminal. Therefore, the (absolute time) duration of time resources (e.g., SF, time slot, or TTI) (collectively referred to as TU (Time Unit) for convenience) consisting of the same number of symbols can be configured differently between the combined cells.

[0085] Figures 3a to 3c is an example diagram illustrating an exemplary architecture for wireless communication services.

[0086] Reference Figure 3a , the UE is connected to the LTE / LTE-A based cell and the NR based cell in DC (dual connectivity) mode.

[0087] The NR-based cell is connected to the core network used for the original fourth-generation mobile communications, namely EPC (Evolved Packet Core).

[0088] Reference Figure 3b , different from Figure 3a , the LTE / LTE-A based cells are connected to a core network for fifth generation mobile communications, i.e., a 5G core network.

[0089] Will be based on Figure 3a and Figure 3b The service mode of the architecture shown is called NSA (non-standalone).

[0090] Reference Figure 3c , the UE is only connected to the NR-based cell. The service mode based on this architecture is called SA (standalone).

[0091] On the other hand, in the NR, it is possible to consider using downlink subframes for reception from the base station and uplink subframes for transmission to the base station. This approach can be applied to both paired and unpaired spectrum. A paired spectrum means including two carrier spectra for downlink and uplink operations. For example, in a paired spectrum, a carrier may include a paired downlink frequency band and an uplink frequency band.

[0092] Figure 4 Shows the time slot structure of the NR frame.

[0093] A slot includes multiple symbols in the time domain. For example, for a normal CP, a slot includes 14 symbols, while for an extended CP, a slot includes 12 symbols. A carrier includes multiple subcarriers in the frequency domain. An RB (Resource Block) is defined as multiple (e.g., 12) consecutive subcarriers in the frequency domain. A BWP (Bandwidth Part) is defined as multiple consecutive (physical, P)RBs in the frequency domain, which can correspond to a parameter set (numerology) (e.g., SCS, CP length, etc.). A terminal can configure up to N (e.g., 4) BWPs in the downlink and uplink, respectively. Downlink or uplink transmission can be performed through an activated BWP, and at a given time, only one BWP established for a terminal can be activated. In a resource grid, each element is called a resource element (RE), which can map a complex symbol.

[0094] Figure 5 Examples of subframe types in NR are shown.

[0095] Figure 5 The TTI (transmission time interval) shown may be referred to as a subframe or time slot for NR (or new RAT). Figure 5 The subframes (or time slots) can be used in the TDD system of NR (or new RAT) to minimize the data transmission delay. Figure 5 As shown, a subframe (or time slot) includes 14 symbols. The first symbols of the subframe (or time slot) can be used for the downlink (DL) control channel, and the last symbols of the subframe (or time slot) can be used for the uplink (UL) control channel. The remaining symbols can be used for DL ​​data transmission or UL data transmission. According to this subframe (or time slot) structure, downlink transmission and uplink transmission can be performed sequentially in one subframe (or time slot). Therefore, downlink data can be received in a subframe (or time slot), and uplink acknowledgment response (ACK / NACK) can also be transmitted in the subframe (or time slot).

[0096] This subframe (or time slot) structure may be referred to as a self-contained subframe (or time slot).

[0097] Specifically, the first N symbols in a time slot can be used to transmit a DL control channel (hereinafter referred to as the DL control region), and the last M symbols in a time slot can be used to transmit a UL control channel (hereinafter referred to as the UL control region). N and M are integers greater than or equal to 0. The resource region between the DL control region and the UL control region (hereinafter referred to as the data region) can be used for DL ​​data transmission or for UL data transmission. For example, in the DL control region, the physical downlink control channel (PDCCH) can be transmitted, and in the DL data region, the physical downlink shared channel (PDSCH) can be transmitted. In the UL control region, the physical uplink control channel (PUCCH) can be transmitted, and in the UL data region, the physical uplink shared channel (PUSCH) can be transmitted.

[0098] Using this subframe (or time slot) structure has the advantage of reducing the time required to retransmit data that has received errors, thereby minimizing the final data transmission standby time. In this self-contained subframe (or time slot) structure, a time gap is required when transitioning from transmit mode to receive mode or vice versa. To this end, a portion of the OFDM symbols when transitioning from DL to UL in the subframe structure can be set as a guard period (GP).

[0099] Figure 6 Shows the structure of a self-contained time slot.

[0100] In the NR system, a frame is characterized by being a self-contained structure that can include all of the DL control channel, DL or UL data, UL control channel, etc. in one time slot. For example, the first N symbols in a time slot can be used to transmit the DL control channel (hereinafter referred to as the DL control region), and the last M symbols in the time slot can be used to transmit the UL control channel (hereinafter referred to as the UL control region). N and M are integers greater than or equal to 0. The resource region between the DL control region and the UL control region (hereinafter referred to as the data region) can be used for DL ​​data transmission or for UL data transmission. As an example, the following structure can be considered. The intervals are listed in chronological order.

[0101] 1.DL only configuration

[0102] 2.UL only configuration

[0103] 3. Mixed UL-DL Configuration

[0104] -DL area + GP (Guard Period) + UL control area

[0105] -DL control area + GP + UL area

[0106] DL area: (i) DL data area, (ii) DL control area + DL data area

[0107] UL area: (i) UL data area, (ii) UL data area + UL control area

[0108] PDCCH can be transmitted in the DL control region, and PDSCH can be transmitted in the DL data region. PUCCH can be transmitted in the UL control region, and PUSCH can be transmitted in the UL data region. DCI (Downlink Control Information), such as DL data scheduling information, UL data scheduling information, etc., can be transmitted in PDCCH. UCI (Uplink Control Information), such as ACK / NACK (Positive Acknowledgement / Negative Acknowledgement) information for DL ​​data, CSI (Channel State Information), SR (Scheduling Request), etc. can be transmitted in PUCCH. GP can provide a time difference in the process of the base station and the terminal switching from the transmission mode to the reception mode or from the reception mode to the transmission mode. Within a subframe, a part of the symbols at the time point of switching from DL to UL can be set as GP.

[0109] <Disclosure of this Specification>

[0110] This specification discloses a terminal that performs beam management using artificial intelligence (AI) / machine learning (ML). Specifically, this specification proposes a solution for a deactivation model for beam failure recovery (BFR) in a cell that performs beam management using an AI / ML model. More specifically, the specification defines the relationship between BFR and model deactivation for a specific cell, and defines a model deactivation event that occurs based on BFR.

[0111] The following describes beam failure detection (BFD) and beam failure recovery (BFR) in the 3GPP NR system.

[0112] Figure 7 Figure 2 shows an example diagram of beam failure detection (BFD) operation in NR.

[0113] BFD is a L1 (Layer 1) / L2 (Layer 2) process in which the PHY (physical) layer provides a beam failure instance (BFI) indication to the MAC (medium access control) layer. Figure 7 The BFI indication from the PHY layer can be provided to the MAC layer based on the BFI indication interval. The MAC layer counts the BFI indications from the PHY layer and declares a beam failure when the counted BFI indications reach a set maximum number.

[0114] Each time the PHY layer detects that the RSRP (reference signal received power) of the reference signal of a serving beam is less than a threshold, it triggers the BFI and provides it to the MAC layer.

[0115] Upon receiving a BFI, the MAC layer starts a timer and continues to increment the counter by 1 for each BFI. That is, if the BFI counter is greater than or equal to the maximum number of BFIs (BFI_COUTER >= beamFailureInstanceMaxCount), the MAC layer triggers a beam failure and initiates the beam failure recovery (BFR) procedure.

[0116] There may be situations where the PHY layer stops providing BFI to the MAC layer, or signal quality improves, causing the PHY layer to no longer detect the problem. To address this situation, the MAC layer increments the BFI counter (BFI_COUTER) by 1 each time the PHY layer reports a BFI and restarts any running timers. If the MAC layer does not receive a BFI from the PHY layer and the (re)started timer expires, the MAC layer can reset the BFI counter (BFI_COUTER) and assume that there is no longer a BFI.

[0117] BFR is a procedure for indicating a new beam (SSB / CSI-RS) when a beam failure is detected for the serving beam, namely the synchronization signal block (SSB) and / or channel state information-reference signal (CSI-RS). Currently, 3GPP NR BFR performs beam failure detection and recovery on a serving cell basis.

[0118] Next, the operations for BFD and BFR defined in the current NR are described.

[0119] 1. The terminal receives BFD / BFR related parameters set for each serving cell through RRC signaling.

[0120] 2. The MAC layer of the terminal receives a BFI indication from the sublayer and starts a BFD timer for the corresponding serving cell.

[0121] 3. If a BFI exceeding the maximum number of BFIs (beamFailureInstanceMaxCount) is received during the running BFD timer, the BFR of the corresponding serving cell will be triggered.

[0122] 4. If the serving cell triggering BFR is a secondary cell (Scell), a BFR MAC CE (control element) is generated and transmitted, or a Scell ​​BFRSR (scheduling request) is triggered.

[0123] 4-A. Include candidate beam information (e.g., SSB / CSI-RS ID) of each cell in the BFR MAC CE and notify the base station of the new beam.

[0124] 4-B. If there is no UL resource for transmitting the BFR MAC CE, resources for transmitting the BFR MAC CE are allocated by transmitting a BFR SR, and then the BFR MAC CE is transmitted.

[0125] 4-C. If it is determined that the BFR MAC CE is successfully transmitted, the BFR procedure is considered to be successfully completed.

[0126] 5. Otherwise, a random access channel (RACH) procedure (contention free random access (CFRA) or contention based random access (CBRA)) is initiated in the SpCell (special cell).

[0127] 5-A. If the RACH triggered by BFR is successfully completed, the BFR procedure is also considered to be successfully completed.

[0128] The following describes technical features related to BFD and BFR procedures. For this description, please refer to 3GPP TS 38.321 v17.3.0 section 5.17.

[0129] When a beam failure is detected by the serving SSB / CSI-RS, a new SSB or CSI-RS can be used to indicate the beam failure recovery procedure of the serving qNB. The MAC entity can be configured by RRC on a per-serving cell or per-beam failure detection-reference signal (BFD-RS) basis. Beam failure is detected by counting BFI indications from the sublayer to the MAC entity.

[0130] RRC sets the following parameters:

[0131] - beamFailureInstanceMaxCount for beam failure detection;

[0132] - beamFailureDetectionTimer for beam failure detection;

[0133] - beamFailureRecoveryTimer for beam failure recovery procedure;

[0134] -rsrp-ThresholdSSB: RSRP threshold for SpCell beam failure recovery;

[0135] -rsrp-ThresholdBFR: RSRP threshold for SpCell beam failure recovery;

[0136] -powerRampingStep: powerRampingStep used for SpCell beam failure recovery;

[0137] -powerRampingStepHighPriority: powerRampingStepHighPriority used for SpCell beam failure recovery;

[0138] -preambleReceivedTargetPower: preambleReceivedTargetPower used for SpCell beam failure recovery;

[0139] -preambleTransMax: preambleTransMax used for SpCell beam failure recovery;

[0140] -scalingFactorBI: scalingFactorBI for SpCell beam failure recovery;

[0141] -ssb-perRACH-Occasion: ssb-perRACHOccasion for SpCell beam failure recovery using contention-free random access resources;

[0142] -ra-ResponseWindow: The time window for monitoring responses to SpCell beam failure recovery using contention-free random access resources;

[0143] -prach-ConfigurationIndex: prach-ConfigurationIndex for SpCell beam failure recovery using contention-free random access resources;

[0144] -ra-ssb-OccasionMaskIndex: ra-ssb-OccasionMaskIndex for SpCell beam failure recovery using contention-free random access resources;

[0145] -ra-OccasionList: ra-OccasionList for SpCell beam failure recovery using contention-free random access resources;

[0146] -candidatesBeamRSList: candidate beam list for SpCell beam failure recovery;

[0147] -candidatesBeamRSSCellList: candidate beam list for SCell beam failure recovery.

[0148] The UE variables used for the BFD procedure are as follows.

[0149] - BFI_COUNTER (per serving cell or per BFD-RS group of a serving cell consisting of two BFD-RS sets): a counter for BFI indication, initially set to 0.

[0150] For each serving cell configured for beam failure detection, the MAC entity performs the following operations:

[0151] >When a BFI indication for a BFD-RS set is received from the lower layer:

[0152] >>Start or restart the beamFailureDetectionTimer of the BFD-RS set;

[0153] >>Increase the BFI_COUNTER of the BFD-RS set by 1;

[0154] >>When BFI_COUNTER of the BFD-RS set is greater than or equal to beamFailureInstanceMaxCount:

[0155] >> Trigger BFR for the BFD-RS set of the serving cell

[0156] >When the serving cell is SpCell and the random access procedure initiated by BFR for two BFD-RS sets of SpCell is successfully completed:

[0157] >>Set the BFI_COUNTER of each BFD-RS set of SpCell to 0.

[0158] >> Consider the beam failure recovery procedure completed successfully.

[0159] >When receiving a BFI indication from the sublayer:

[0160] >>Start or restart beamFailureDetectionTimer;

[0161] >>Increment BFI_COUNTER by 1;

[0162] >>BFI_COUNTER>=beamFailureInstanceMaxCount:

[0163] >>>If the serving cell is SCell:

[0164] >>>>Trigger BFR for this serving cell

[0165] >>> Otherwise:

[0166] >>>>Start the random access procedure on SpCell.

[0167] Currently, 3GPP has started research on applying AI / ML to beam management from Release 18 to improve the performance of beam management and reduce the burden of beam measurement.

[0168] The following table discusses the list of terms applicable to AI / ML.

[0169]

Table 5

[0170]

[0171]

[0172]

[0173]

[0174] 3GPP is studying the impact of beam management (BM) procedures on the specifications of BM Case 1 (spatial beam prediction) and BM Case 2 (temporal beam prediction) using AI / ML models.

[0175] Furthermore, model management procedures will be introduced through model performance evaluation. Model management, a process for managing model performance, such as model updates, model exchanges, model activation / deactivation, and fallback, is expected to be discussed in detail. Model management can be performed through model monitoring, and it is expected that model monitoring procedures will define different methods depending on whether the model is a user-side (UE-side) or a network-side (NW-side) model.

[0176] At the same time, the beam management operation in traditional NR increases with the number of beams and the number of terminals, leading to problems of increased system overhead and terminal power consumption. In addition, for terminals in the initial access stage of the cell, the terminal has to go through the process of selecting the initial beam after measuring all beams, which may cause delays in cell access. In order to improve this problem, it is recommended to use an AI / ML model that predicts the strength of the entire beam by measuring part of the beam, but the detailed procedures for this or the content of the scheme have not yet been defined. In this specification, it is recommended to use a model management scheme based on the occurrence of BFR when the AI / ML model is applied to the beam management procedure.

[0177] Figures 8a to 8b Beam measurement and spatial domain beam prediction using AI / ML are shown.

[0178] Currently, 3GPP RAN (radio access network) WG1 (working group 1) has started research on "AI / ML for beam management" and agreed to discuss spatial DL beam prediction (BM-Case1) and temporal DL beam prediction (BM-Case2) as sub-use cases. This predicts the strength of the beams of set A by measuring the beams belonging to set B. The case of spatial DL beam prediction is shown in Figures 8A to 8B, and Figure 8A shows the case where set B is a subset of set A, and Figure 8B considers that set B is composed of wide beams and set A is composed of narrow beams, that is, sets composed of different beams. For temporal DL beam prediction, in addition to the cases of i) Set B being a subset of Set A and ii) Set A and Set B being different sets for spatial DL beam prediction, the case of iii) Set A and Set B being the same set can be considered. Temporal DL beam prediction predicts future beam information based on past beam measurement information. Therefore, a solution can be considered that predicts the entire beam based on spatial DL beam prediction and then applies this solution to the case of iii) Set A and Set B being the same set. For this reason, the cases of i) Set B being a subset of Set A and ii) Set A and Set B being different sets for spatial DL beam prediction are expected to be used as the basic beam prediction schemes.

[0179] Figure 9 Shown is temporal domain beam prediction using AI / ML.

[0180] Temporal beam prediction in BM-Case 2 is defined as the operation of predicting the beam result (i.e., output) at a specific time point in the near future based on the past beam measurement result information (i.e., input), as follows: Figure 9 At this time, in addition to the above-mentioned case i) that set B is a subset of set A and ii) that set A and set B are different sets, the beam used for input and the beam set derived as output can also consider the case iii) that set A and set B are composed of the same set.

[0181] Meanwhile, AI / ML-based beam prediction can be defined differently depending on the location of the AI / ML model and the subject of training / inference. The following are alternatives.

[0182] Alternative 1: Training and inference of AI / ML models on the network side

[0183] Alternative 2: Training and inference of AI / ML models on the UE side

[0184] -Alternative 3: Training AI / ML models on the network side and inferring AI / ML models on the UE side

[0185] -Alternative 4: Training AI / ML models on the UE side and inferring AI / ML models on the network side

[0186] In the above alternatives, for Alternatives 1 and 2, both training and inference run on one node, so model transfer on the air interface is not required, but the information required to operate the AI / ML model must be signaled. This may belong to the currently defined collaboration level y (no model transfer but only signaling). For Alternatives 3 and 4, training and inference run on different nodes, so model transfer on the air interface is required. This may belong to the currently defined collaboration level z. Currently, 3GPP has decided to support Alternatives 1 and 2, but has decided to postpone support for Alternative 3 until it is decided whether to support model transfer of UE-side AI / ML models. In addition, Alternative 4 has decided not to support beam management in release 18.

[0187] In addition, the beam management method using AI / ML is currently being considered. In order to predict the signal strength of all beams (set A), some beams in all beams or all beams are measured with other beams, expecting to reduce the burden of all beam measurements and reduce the overhead of reference signals (RS). However, if the predicted value of the beam of the predicted set (Set) A is inaccurate, the AI / ML method is used to select inaccurate beams, resulting in frequent beam failures or continuous disconnection. In order to solve this problem, it is necessary to continuously monitor the accuracy of the predicted beams. Currently, 3GPP has decided to conduct research on such monitoring. Depending on the node that calculates the model performance and the node that determines the model management, the following three model monitoring methods can be considered.

[0188] -Alternative 1: UE-side model monitoring

[0189] >UE monitoring performance metric

[0190] >UE determines model selection / activation / deactivation / switching / fallback operations

[0191] -Alternative 2: Network-side model monitoring

[0192] >Network monitoring performance metrics

[0193] >Network determines model selection / activation / deactivation / switching / fallback operations

[0194] -Alternative 3: Hybrid model monitoring

[0195] >UE monitoring performance metric

[0196] >Network determines model selection / activation / deactivation / switching / fallback operations

[0197] However, for the case of UE-side AI / ML models, the three monitoring alternatives mentioned above include UE-side model monitoring, network-side model monitoring, and hybrid model monitoring. However, for the case of NW-side AI / ML models, it was decided to only consider NW-side model monitoring for model monitoring in alternative 2.

[0198] Here, based on the results of the monitoring performance evaluation, the model can be selected, activated, deactivated, switched, or fallbacked by the terminal (UE) or the network (NW). In addition to model accuracy, system performance is also being considered as a metric for model performance evaluation. In the case of beam management, frequent BFR can reduce system performance, but this has not yet been discussed.

[0199] AI / ML-based beam management methods are considering schemes for updating and converting models through model monitoring to prevent frequent beam connection failures due to such inaccurate beam inference. However, due to sudden changes in the model's environment, there is a possibility that model performance improvement is in a difficult environment. From this perspective, the content described in this specification recommends considering the occurrence of BFR as one of the performance evaluation metrics of the AI / ML model for beam management, and recommends performing model deactivation or fallback operations according to the occurrence of BFR to manage the model.

[0200] In addition, this specification recommends that when executing the beam management procedure between the base station (gNB / NW) and the terminal, the corresponding AI / ML model be deactivated (or fallbacked) according to the BFR that occurs in the cell where beam management is performed using the AI / ML model. In other words, it is recommended to define an event for model deactivation based on the number of BFR occurrences. As a specific solution for this, an arbitrary time interval for determining model deactivation (model deactivation timer: Model_deact_Timer) and new parameters for counting the number of BFRs that occur during the corresponding time period (BFR number and BFR maximum number: BFR_num and Max_Num ofBFR) are defined, and terminal operations utilizing them are defined.

[0201] Figure 10 A method for operating a terminal according to an embodiment of this specification is shown.

[0202] refer to Figure 10 The terminal triggers beam failure recovery (BFR) for a cell that uses an artificial intelligence and / or machine learning (AI / ML) model to perform beam management (S1001). Thereafter, based on the number of BFRs triggered within a specific time interval, the terminal indicates deactivation of the artificial intelligence and / or machine learning model (S1002).

[0203] That is, when BFR is triggered for a cell corresponding to any AI / ML model activated in the terminal, an instruction is given to deactivate the AI / ML model if the number of BFRs occurring within any time interval of the corresponding model meets a defined event.

[0204] In this specification, for the terminal side model (UE-side model), the terminal side model monitoring (UE-side model monitoring) or hybrid model monitoring (hybrid model monitoring) is preferably applied. However, even in the case of the network side model (NW-side model), the base station can set corresponding timers and parameters to the terminal for BFR-related model deactivation (model deactivation). This can receive BFR-related information from the terminal to determine model deactivation. Alternatively, as an internal implementation of the base station, the corresponding timer or parameter is set, so that the execution model deactivation can also be defined.

[0205] More specifically, the model deactivation events suggested in this specification, through timers and BFR count parameters, can be defined as any of i) model, ii) serving cell, or iii) terminal operation activated for the terminal. Thus, the method for setting the values ​​of the corresponding timers and parameters, whether related to or independent of the alternatives, can be set by the network per terminal, per model, or per serving cell.

[0206] Figures 11a to 11b An example of model deactivation based on model execution according to an embodiment of this specification is shown.

[0207] When more than one serving cell is set for any terminal and any ML model for beam management is applicable to more than one serving cell, a timer and a BFR number parameter can be used to count the number of BFRs that occur in the cells to which the model is applied.

[0208] Figure 11a Shows the case where both events and deactivations are performed based on the model, Figure 11b It shows the case where events are executed based on the model and model deactivation is executed based on the terminal.

[0209] refer to Figure 11a If three serving cells are configured for a terminal, and beam management is performed using AI / ML model X for serving cell 1, and AI / ML model Y for the remaining serving cells 2 and 3, then the timer (X_timer) and BFR count parameter (X_COUNT) for BFR operations in serving cell 1 using AI / ML model X, as well as the timer (Y_timer) and BFR count parameter (Y_COUNT) for BFR operations in serving cells 2 and 3 using AI / ML model Y, can be managed / operated based on models X and Y. When BFR occurs in serving cell 2 or 3, the timer for model Y is started, and additional BFR occurrences in serving cell 2 or 3 are counted during the timer's operation. For models whose BFR count reaches the maximum, deactivation is indicated. That is, if model Y used in serving cell 2 or 3 is deactivated, beam management for serving cell 2 or 3 can fall back to the legacy method.

[0210] At the same time, Figure 11b In the timer and BFR frequency parameters are as follows Figure 11aThe models are run identically, but if the number of BFRs reaches the maximum value during the timer for a specific model, all AI / ML models related to beam management activated for the terminal are deactivated. This means that the reliability of model Y is considered to be the same as that of all beam management models, so the terminal discontinues the beam management process using the AI / ML model and falls back to the traditional beam management method.

[0211] Figures 12a to 12c An example of model deactivation performed based on a serving cell according to an embodiment of this specification is shown.

[0212] When one or more serving cells are set for any terminal and AI / ML model X and model Y for beam management are applied to serving cell 1, serving cell 2, and serving cell 3, respectively, even if the same model Y is applied and operated in serving cell 2 and serving cell 3, the model deactivation timers (C1_timer, C2_timer, C3_timer) and BFR count parameters (C1_COUNT, C2_COUNT, C3_COUNT) can be defined / operated for each serving cell.

[0213] Figure 12a Shows the case where both event and model deactivation are performed based on the serving cell, Figure 12b shows the case where events are performed based on serving cells and model deactivation is performed based on models, Figure 12c This shows a case where an event is performed based on a serving cell and model deactivation is performed based on a terminal.

[0214] refer to Figure 12a , a timer is started separately for the serving cell where BFR is triggered, and during the operation of the corresponding timer, only the BFRs that occur in the corresponding cell are counted as valid times. If the number of BFRs that occur in the corresponding cell reaches the maximum value, the beam management for the corresponding cell can be instructed (deactivated or rolled back) to not apply the AI / ML model. In other words, even if one model Y is applied to serving cell 2 and serving cell 3 and beam management is performed, if BFR occurs frequently only in serving cell 2, the beam management applying model Y can be deactivated only for serving cell 2, and the beam management applying model Y can continue to be performed for serving cell 3.

[0215] At the same time, if Figure 12b As shown, events are run on a cell-by-cell basis, but deactivation can also be applied to specific models. Figure 12aFor example, timers (C1_timer, C2_timer, C3_timer) and BFR count parameters (C1_COUNT, C2_COUNT, C3_COUNT) are applied to each cell and run independently. However, if the BFR count for serving cell 2 reaches a maximum value during the operation of the timer for serving cell 2, the AI / ML model for all cells using model Y may be deactivated. In other words, beam management using AI / ML may be deactivated for serving cell 2 and serving cell 3 to which model Y is applied, and a fallback to the conventional beam management procedure may be performed for serving cell 2 and serving cell 3.

[0216] On the other hand, Figure 12c As shown, the timers (C1_timer, C2_timer, C3_timer) and BFR count parameters (C1_COUNT, C2_COUNT, C3_COUNT) run for each serving cell, but if the number of BFRs reaches the maximum value during the operation of the timer for a specific cell, it can be defined as deactivating all beam management-related AI / ML models activated for the terminal.

[0217] Figure 13 An example of model deactivation based on terminal execution according to an embodiment of this specification is shown.

[0218] refer to Figure 13 , a timer and a BFR number parameter are defined for the terminal, and no matter in which cell the BFR occurs in the terminal, a timer (UE_timer) and a parameter (UE_COUNT) are used to count the BFR number. Figure 13 In the example of FIG, if the number of BFRs reaches a maximum value during the operation of the timer, all AI / ML models for beam management activated for the terminal can be deactivated.

[0219] In this specification, the above content can be applied to the beam management of the AI / ML model applicable to the serving cell. Since the current BFR is defined to operate in units of serving cells, the scheme for determining the event of deactivation of the BFR model is preferably as follows: Figures 12a to 12c An example of this is run based on the serving cell.

[0220] Meanwhile, if the above content in this specification is used to judge the reliability of the model, it is preferred to perform an operation of deactivating the corresponding model (or beam management using the AI / ML model) applicable to the terminal according to the number of consecutive BFRs of a specific cell (for example, n times or more, where n is an integer greater than or equal to 1). That is, if Figures 12b to 12cIn the example of , it is preferred to deactivate "Beam management using AI / ML model" for (all) service cells that perform beam management using the corresponding AI / ML model, and perform the operation of falling back to the traditional beam management procedure.

[0221] If the maximum number of BFRs mentioned above in this specification is greater than 2, it is preferred to set the length of the timer to a sufficiently long time value, considering the time required to determine the success / failure of BFR after triggering BFR and the time required to declare a new BFR (receive continuous BFI) thereafter.

[0222] The operations related to model deactivation described above in this specification are preferably performed at the MAC layer.

[0223] In addition, matters described in this specification are applicable to both the terminal side model (UE-side model) and the network side model (NW-side model).

[0224] <Order of Examples of This Specification>

[0225] Figure 14 An example of operations of a base station and a terminal according to an embodiment of this specification is shown.

[0226] refer to Figure 14 , the base station transmits a message containing setting information related to the AI / ML model for beam management to the terminal, and the terminal receives the message (S1401). The message containing setting information related to the AI / ML model can be an RRC message for AI / ML model management, or an RRC message for beam management based on the AI / ML model. The setting information related to the AI / ML model may include at least one of the following: i) AI / ML model information for the beam management procedure; ii) timer time value information related to the AI / ML model (model deactivation timer: model_DeactivationTimer); and iii) BFR maximum number information related to the timer (model BFR maximum number: model_bfrMax).

[0227] The terminal performs AI / ML model-based beam management (S1402), which can be based on the configuration information related to the AI / ML model transmitted by the base station to the terminal. In other words, it can be based on the model deactivation timer (model_DeactivationTimer) information and the model BFR maximum number (model_bfrMax) information applicable to each serving cell. In addition, AI / ML model-based beam management is performed using the UE variable BFR counter (BFR_COUNTER / BFR_CNT) for the BFR procedure applicable to each serving cell.

[0228] When the terminal triggers BFR for a serving cell that applies an AI / ML model, if there is no timer (model_DeactivationTimer) running, the terminal starts a timer (model_DeactivationTimer). Then, while the timer (model_DeactivationTimer) is running, the terminal performs BFR triggered by BFR (S1403 to S1404). BFR may include RACH procedures, SR transmissions, and / or MAC CE transmissions.

[0229] Based on BFR triggering, the terminal increments the BFR counter (BFR_CNT) by 1. If the BFR counter (BFR_CNT) reaches the model BFR maximum number (model_bfrMax), that is, the BFR counter (BFR_CNT) is equal to or greater than the model BFR maximum number (model_bfrMax), the timer (model_DeactivationTimer) is stopped. Then, the corresponding AI / ML model is deactivated (S1405), and the deactivation of the corresponding AI / ML model is notified to the RRC layer of the terminal.

[0230] When the timer (model_DeactivationTimer) expires, the counter (BFR_CNT) is set to 0.

[0231] At the same time, if it is a network management model or a network-side model (NW-side model), the terminal notifies the base station of the deactivation of the corresponding AI / ML model. That is, the terminal transmits the deactivation information of the corresponding AI / ML model to the base station, and the base station receives the information (S1406). The deactivation information of the AI / ML model may include at least one of the following: i) ID (identity) information of the deactivated AI / ML model; ii) an indicator indicating that it is to be deactivated; and iii) an indicator notifying of deactivation. Among them, if it is a network-side model (NW-side model), it may include ii) an indicator indicating that it is to be deactivated; if model management (model management) is performed in the network, it may include iii) an indicator notifying of deactivation.

[0232] If the base station receives deactivation information of the AI / ML model from the terminal, the beamforming related to the AI / ML model applicable to the beam management of the terminal can be interrupted based on this, and fallback to the traditional beam management technology. If it is a network-side model (NW-side model), the AI / ML model in the network, that is, running in the base station, is deactivated (S1407). If the network, that is, performs model management in the base station, then the model management related to the corresponding AI / ML model of the terminal is performed.

[0233] Figure 15 A method for operating a terminal according to an embodiment of this specification is shown.

[0234] refer to Figure 15 The terminal triggers at least one beam failure recovery (BFR) for a cell that performs beam management using an artificial intelligence and / or machine learning (AI / ML) model (S1501). Then, based on the number of at least one beam failure recovery triggered within a specific time period, the terminal deactivates the artificial intelligence and / or machine learning model (S1502). Thereafter, the terminal transmits deactivation information of the artificial intelligence and / or machine learning model to a base station (S1503).

[0235] The terminal may start a timer associated with the deactivation of the artificial intelligence and / or machine learning model, and when the started timer is running, it may count the number of times the triggered at least one beam failure is recovered.

[0236] In addition, when the counted number of at least one beam failure recovery reaches a maximum number, the terminal may stop the timer. In addition, when the timer expires, the terminal may initialize the counted number of at least one beam failure recovery.

[0237] At the same time, the terminal can receive setting information related to the artificial intelligence and / or machine learning model from the base station, but the setting information related to the artificial intelligence and / or machine learning model may include at least one of timer value information and maximum number of beam failure recovery information.

[0238] In addition, the deactivation information transmitted by the terminal to the base station may include at least one of the following: i) ID (identity) information of the artificial intelligence and / or machine learning model, ii) information notifying that the artificial intelligence and / or machine learning model is deactivated, and iii) an indicator indicating the deactivation of the artificial intelligence and / or machine learning model of the base station.

[0239] Figure 16The following describes an operating method of a base station according to an embodiment of the present specification.

[0240] refer to Figure 16 , the base station transmits setting information related to the artificial intelligence and / or machine learning model to the terminal (S1601). Then, the base station receives deactivation information of the artificial intelligence and / or machine learning model from the terminal (S1602). Thereafter, based on the received deactivation information, the base station interrupts beamforming related to the artificial intelligence and / or machine learning model (S1603).

[0241] At the same time, the base station can send setting information related to the artificial intelligence and / or machine learning model to the terminal, but the setting information related to the artificial intelligence and / or machine learning model can include at least one of timer value information and maximum number of beam failure recovery information.

[0242] In addition, the deactivation information received by the base station from the terminal may include at least one of the following: i) ID (identity) information of the artificial intelligence and / or machine learning model, ii) information notifying that the artificial intelligence and / or machine learning model is deactivated, and iii) an indicator indicating the deactivation of the artificial intelligence and / or machine learning model of the base station.

[0243] On the one hand, based on the deactivation information received from the terminal, the base station can fallback to beam management that does not apply artificial intelligence and / or machine learning models.

[0244] The contents disclosed in this specification can be applied independently or in any combination. In addition, this specification is based on the 5G NR system, but is not related to specific wireless communication technologies. All situations where the concepts of this specification are applicable are included in the scope of this specification.

[0245] Figure 17 A device according to an embodiment of the present specification is shown.

[0246] Reference Figure 17 , a wireless communication system may include a first device 100a and a second device 100b.

[0247] The first device 100a can be a base station, a network node, a transmitting terminal, a receiving terminal, a wireless device, a wireless communication device, a vehicle, a vehicle equipped with an autonomous driving function, a smart car (Connected Car), an unmanned aerial vehicle (UAV), an AI (Artificial Intelligence) module, a robot, an AR (Augmented Reality) device, a VR (Virtual Reality) device, an MR (Mixed Reality) device, a holographic device, a public safety device, an MTC device, an IoT device, a medical device, a fintech device (or financial device), a security device, a climate / environmental device, a 5G service-related device or other devices related to the fourth industrial revolution.

[0248] The second device 100b can be a base station, a network node, a transmitting terminal, a receiving terminal, a wireless device, a wireless communication device, a vehicle, a vehicle equipped with an autonomous driving function, a smart car (Connected Car), an unmanned aerial vehicle (UAV), an AI (Artificial Intelligence) module, a robot, an AR (Augmented Reality) device, a VR (Virtual Reality) device, an MR (Mixed Reality) device, a holographic device, a public safety device, an MTC device, an IoT device, a medical device, a fintech device (or financial device), a security device, a climate / environmental device, a 5G service-related device or other devices related to the fourth industrial revolution.

[0249] The first device 100a may include: at least one processor, such as a processor 1020a; at least one memory, such as a memory 1010a; and at least one transceiver, such as a transceiver 1031a. The processor 1020a may perform the aforementioned functions, steps, and / or methods. The processor 1020a may execute one or more protocols. For example, the processor 1020a may execute one or more layers of a wireless interface protocol. The memory 1010a may be connected to the processor 1020a and store various forms of information and / or instructions. The transceiver 1031a may be connected to the processor 1020a and control the transmission and reception of wireless signals.

[0250] The second device 100b may include at least one processor, such as a processor 1020b, at least one memory device, such as a memory 1010b, and at least one transceiver, such as a transceiver 1031b. The processor 1020b may perform the aforementioned functions, steps, and / or methods. The processor 1020b may implement one or more protocols. For example, the processor 1020b may implement one or more layers of a wireless interface protocol. The memory 1010b may be connected to the processor 1020b and store various forms of information and / or instructions. The transceiver 1031b may be connected to the processor 1020b and control the transmission and reception of wireless signals.

[0251] The memory 1010a and / or the memory 1010b may be connected to the inside or outside of the processor 1020a and / or the processor 1020b, respectively, or may be connected to other processors via various technologies such as wired or wireless connections.

[0252] The first device 100a and / or the second device 100b may have more than one antenna. For example, the antenna 1036a and / or the antenna 1036b may be configured to transmit and receive wireless signals.

[0253] Figure 18 It is a block diagram showing the structure of a terminal according to an embodiment of this specification.

[0254] in particular Figure 18 The foregoing is shown in more detail Figure 17 Figure 1.

[0255] The device includes a memory 1010 , a processor 1020 , a transceiver 1031 , a power management module 1091 , a battery 1092 , a display 1041 , an input unit 1053 , a speaker 1042 , a microphone 1052 , a SIM (subscriber identification module) card, and one or more antennas.

[0256] The processor 1020 may be configured to implement the functions, steps and / or methods described and proposed in this specification. The layers of the radio interface protocol may be implemented in the processor 1020. The processor 1020 may include an ASIC (application-specific integrated circuit), other chipsets, logic circuits and / or data processing devices. The processor 1020 may be an AP (application processor). The processor 1020 may include at least one of a DSP (digital signal processor), a CPU (central processing unit), a GPU (graphics processing unit), and a modem (modulator and demodulator). The processor 1020 may be, for example, SNAPDRAGONTM series processors manufactured by EXYNOSTM series processors manufactured by A series processors manufactured by HELIOTM series processors manufactured by ATOMTM series processors manufactured by KIRINTM series processors or corresponding new generation processors manufactured by.

[0257] The power management module 1091 manages power to the processor 1020 and / or the transceiver unit 1031. The battery 1092 supplies power to the power management module 1091. The display device 1041 outputs results processed by the processor 1020. The input unit 1053 receives input to be used by the processor 1020. The input unit 1053 can be displayed on the display device 1041. A SIM card is an integrated circuit used to identify and authenticate a subscriber's IMSI (International Mobile Subscriber Identity) in mobile devices such as cell phones and computers, as well as to securely store cryptographic keys associated with the IMSI. Many SIM cards also store phonebook information.

[0258] The memory 1010 is operably combined with the processor 1020 to store various information for operating the processor 610. The memory 1010 may include ROM (read-only memory), RAM (random access memory), flash memory, memory card, storage medium and / or other storage devices. When the embodiment is implemented in software, the technology described in this specification can be implemented in modules (e.g., steps, functions, etc.) that perform the functions described in this specification. The modules can be stored in the memory 1010 and run by the processor 1020. The memory 1010 can be implemented inside the processor 1020. Alternatively, the memory 1010 can be implemented outside the processor 1020 and can be communicatively connected to the processor 1020 by various means known in the art.

[0259] The transceiver 1031 is operably coupled to the processor 1020 to transmit and / or receive wireless signals. The transceiver 1031 includes a transmitter and a receiver. The transceiver 1031 may include baseband circuitry for processing radio frequency signals. The transceiver controls one or more wires to transmit and / or receive wireless signals. To initiate communication, the processor 1020 sends instruction information to the transceiver 1031 to transmit wireless signals, such as those constituting voice communication data. The antenna performs the functions of transmitting and receiving wireless signals. When receiving wireless signals, the transceiver 1031 may send the signals and convert the signals to baseband for processing by the processor 1020. The processed signals may be converted into auditory or visual information output through the speaker 1042.

[0260] The speaker 1042 outputs sound-related results processed by the processor 1020. The microphone 1052 receives sound-related input to be used by the processor 1020.

[0261] The user inputs command information such as a phone number by, for example, pressing a button (or touching) input unit 1053 or by voice activation via microphone 1052. Processor 1020 receives this command information and processes it to execute appropriate functions, such as dialing the phone number. Operational data can be retrieved from the SIM card or memory 1010. Processor 1020 can also display the command information or operational information on display device 1041 for user convenience.

[0262] Figure 19 A block diagram showing the configuration of a processor that implements the disclosure of this specification.

[0263] Reference Figure 19It is understood that the processor 1020 implementing the disclosures of this specification may include multiple circuits to implement the functions, steps, and / or methods described and proposed in this specification. For example, the processor 1020 may include a first circuit 1020-1, a second circuit 1020-2, and a third circuit 1020-3. In addition, although not shown, the processor 1020 may include more circuits. Each circuit may include multiple transistors.

[0264] The processor 1020 may also be referred to as an ASIC (application-specific integrated circuit) or an AP (application processor), and may include at least one of a DSP (digital signal processor), a CPU (central processing unit), and a GPU (graphics processing unit).

[0265] Figure 20 To show in detail Figure 17 The transceiver of the first device or Figure 18 Block diagram of the transceiver portion of the device shown.

[0266] Reference Figure 20 The transceiver unit 1031 includes a transmitter 1031-1 and a receiver 1031-2. The transmitter 1031-1 includes a DFT (Discrete Fourier Transform) unit 1031-11, a subcarrier mapper 1031-12, an IFFT unit 1031-13, a CP insertion unit 1031-14, and a wireless transmission unit 1031-15. The transmitter 1031-1 may also include a modulator. In addition, for example, a scrambler (not shown), a modulation mapper (not shown), a layer mapper (not shown), and a layer permutator (not shown) may be further included, which may be configured before the DFT unit 1031-11. That is, in order to prevent an increase in PAPR (peak-to-average power ratio), the transmitter 1031-1 first passes the information through the DFT 1031-11 before mapping the signal to the subcarriers. The signal spread (or equivalently precoded) by the DFT unit 1031 - 11 is subjected to subcarrier mapping by the subcarrier mapper 1031 - 12 and then passes through the IFFT (Inverse Fast Fourier Transform) unit 1031 - 13 to form a signal on the time axis.

[0267] The DFT unit 1031-11 performs DFT on the input symbols to output complex symbols (complex-valued symbols). For example, if Ntx symbols are input (but Ntx is a natural number), the DFT size is Ntx. The DFT unit 1031-11 can be called a transform precoder. The subcarrier mapper 1031-12 maps the complex symbols to each subcarrier in the frequency domain. The complex symbols can be mapped to resource elements corresponding to resource blocks allocated for data transmission. The subcarrier mapper 1031-12 can be called a resource element mapper. The IFFT unit 1031-13 can perform IFFT on the input symbols to output a baseband signal for data as a time domain signal. The CP insertion unit 1031-14 copies the latter part of the baseband signal for data and inserts it into the front part of the baseband signal for data. By inserting CP, ISI (Inter-Symbol Interference) and ICI (Inter-Carrier Interference) can be prevented, and orthogonality can be maintained in multiplexed channels.

[0268] On the other hand, receiver 1031-2 includes a wireless receiving unit 1031-21, a CP removal unit 1031-22, an FFT unit 1031-23, and an equalizer 1031-24. The wireless receiving unit 1031-21, CP removal unit 1031-22, and FFT unit 1031-23 of receiver 1031-2 perform the opposite functions of the wireless transmitting unit 1031-15, CP insertion unit 1031-14, and IFF unit 1031-13 of transmitter 1031-1. Receiver 1031-2 may also include a demodulator.

[0269] The preferred embodiments are described above by way of example, but the disclosure of this specification is not limited to these specific embodiments, and thus the present invention may be modified, altered, or improved into various forms within the scope of the concept of this specification and the claims.

[0270] In the exemplary system described above, the method is described as a series of steps or blocks based on the flowchart, but the steps are not limited to the order described. Some steps may be performed in a different order than described or simultaneously. In addition, those skilled in the art will understand that the steps shown in the flowchart are not exclusive. Other steps may be included or one or more steps in the flowchart may be deleted without affecting the scope of the rights.

[0271] The claims described in this specification may be combined in various ways. For example, the technical features of method claims described in this specification may be combined and implemented by a device, and the technical features of device claims described in this specification may be combined and implemented by a method. Furthermore, the technical features of method claims and device claims described in this specification may be combined and implemented by a device, and the technical features of method claims and device claims described in this specification may be combined and implemented by a method.

Claims

1. A method, as a method for performing beam management by a terminal in a wireless communication system, wherein: The method comprises: For cells that utilize artificial intelligence and / or machine learning models to perform beam management, trigger at least one step of beam failure recovery; Deactivating the artificial intelligence and / or machine learning model based on the number of times at least one beam failure is recovered by the triggering within a specific time interval; and The step of sending deactivation information of the artificial intelligence and / or machine learning model to a base station.

2. The method according to claim 1, wherein The method further comprises: the step of starting a timer associated with deactivation of said artificial intelligence and / or machine learning model; and During the driving of the started timer, the number of times the triggered at least one beam failure is recovered is counted.

3. The method according to claim 2, wherein: The method further comprises: The step of stopping the timer when the counted number of at least one beam failure recovery reaches a maximum number.

4. The method according to claim 2, wherein: The method further comprises: When the timer expires, the step of initializing the counting of the number of times at least one beam failure is restored.

5. The method according to claim 1, wherein The method further comprises: The step of receiving setting information related to the artificial intelligence and / or machine learning model.

6. The method according to claim 5, wherein: The setting information related to the artificial intelligence and / or machine learning model includes at least one of timer value information and maximum number of beam failure recovery times.

7. The method according to claim 1, wherein The deactivation information includes at least one of the following: i) identity information of the artificial intelligence and / or machine learning model; ii) information notifying that the artificial intelligence and / or machine learning model is deactivated; and iii) an instruction instructing the deactivation of the artificial intelligence and / or machine learning model of the base station.

8. A method for performing beam management by a base station in a wireless communication system, wherein: The method comprises: The step of sending setting information related to the artificial intelligence and / or machine learning model to the terminal; A step of receiving deactivation information of the artificial intelligence and / or machine learning model from the terminal; and Based on the received deactivation information, a beam forming step associated with the artificial intelligence and / or machine learning model is interrupted.

9. The method according to claim 8, wherein The setting information related to the artificial intelligence and / or machine learning model includes at least one of timer value information and maximum number of beam failure recovery times.

10. The method according to claim 8, wherein The deactivation information includes at least one of the following: i) identity information of the artificial intelligence and / or machine learning model; ii) information notifying that the artificial intelligence and / or machine learning model is deactivated; and iii) an instruction instructing the deactivation of the artificial intelligence and / or machine learning model of the base station.

11. The method according to claim 8, wherein The method further comprises: Based on the received deactivation information, fall back to a step of not applying beam management of the artificial intelligence and / or machine learning model.

12. A communication device, as a communication device in a wireless communication system, comprising: at least one processor; as well as at least one memory storing instructions and electrically operatively connected to the at least one processor, and Based on the instruction being executed by the at least one processor, the operations performed include: For cells that utilize artificial intelligence and / or machine learning models to perform beam management, trigger at least one step of beam failure recovery; performing a step of deactivating the artificial intelligence and / or machine learning model based on the number of times the at least one beam failure recovery is triggered within a specific time interval; and The step of sending deactivation information of the artificial intelligence and / or machine learning model to a base station.

13. The communication device according to claim 12, wherein: Based on the instruction being executed by the at least one processor, the executed operations further include: the step of starting a timer associated with deactivation of said artificial intelligence and / or machine learning model; and During the driving of the started timer, the number of times the triggered at least one beam failure is recovered is counted.

14. The communication device according to claim 13, wherein: Based on the instruction being executed by the at least one processor, the executed operations further include: The step of stopping the timer when the counted number of at least one beam failure recovery reaches a maximum number.

15. The communication device according to claim 13, wherein: Based on the instruction being executed by the at least one processor, the executed operations further include: When the timer expires, the step of initializing the counting of the number of times at least one beam failure is restored.

16. The communication device according to claim 12, wherein: Based on the instruction being executed by the at least one processor, the executed operations further include: The step of receiving setting information related to the artificial intelligence and / or machine learning model.

17. The communication device according to claim 16, wherein: The setting information related to the artificial intelligence and / or machine learning model includes at least one of timer value information and maximum number of beam failure recovery times.

18. The communication device according to claim 12, wherein: The deactivation information includes at least one of the following: i) identity information of the artificial intelligence and / or machine learning model; ii) information notifying that the artificial intelligence and / or machine learning model is deactivated; and iii) an instruction instructing the deactivation of the artificial intelligence and / or machine learning model of the base station.

Images