Techniques for joint channel state information training and precoder matrix indicator feedback for artificial intelligence-enabled networks
By separating and processing two subsets in the CSI report, the problem of the CSI feedback compression method limiting the design capabilities of AI/ML models in the prior art is solved, and more flexible response to environmental changes and improved robustness are achieved.
Patent Information
- Application Number
- CN202380030706.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2022-04-15
- Filing Date
- 2023-04-17
- Publication Date
- 2024-11-15
AI Technical Summary
In prior art In wireless communication systems that enable artificial intelligence/machine learning (AI/ML) models, the compression method of CSI feedback limits the design capabilities of the AI/ML model, and conventional CSI feedback schemes do not match the channel allocation for environmental changes. Inflexible enough to affect the robustness of the model.
A mechanism is proposed to support AI/ML model training by separating two subsets in CSI reports, one reported in the compressed/transformed domain, to reduce feedback overhead and the other reported in the uncompressed/transformed domain. , and select the appropriate CSI compression type on the network side to deal with environmental changes.
It effectively reduces the overhead of CSI feedback, and provides a more flexible way to deal with environmental changes, improving the robustness of AI/ML models and the quality of training data.
Smart Images

Figure CN118975147A_ABST
Abstract
Description
Technical Field
[0001] Generally, the subject matter disclosed herein relates to wireless communications, and more particularly, to techniques for joint channel state information ("CSI") training and precoder matrix indicator ("PMI") feedback for artificial intelligence ("AI") enabled networks. Background Art
[0002] A wireless communication system may include one or more network communication devices (e.g., base stations), which may be further referred to as eNodeBs ("eNBs"), new generation Node Bs ("gNBs"), or other suitable terms. Each network communication device (e.g., base station) may support wireless communication for one or more user communication devices, which may be further referred to as user equipment ("UE") or other suitable terms. A wireless communication system may support wireless communication with one or more user communication devices by utilizing resources of the wireless communication system (e.g., time resources (e.g., symbols, time slots, subframes, frames, etc.) or frequency resources (e.g., subcarriers, carriers)). In addition, a wireless communication system may support wireless communication across various radio access technologies, including third generation ("3G") radio access technologies, fourth generation ("4G") radio access technologies, fifth generation ("5G") radio access technologies, and other suitable radio access technologies (e.g., sixth generation ("6G")) other than 5G. In a wireless communication system, one or more of a network communication device (e.g., a base station) or a user communication device (e.g., a UE) may support one or more CG configurations for wireless communication (e.g., downlink communication, uplink communication). Summary of the invention
[0003] A solution of techniques for joint CSI training and PMI feedback for AI-enabled networks is disclosed.
[0004] In one embodiment, a first device includes a processor and a memory coupled to the processor. In one embodiment, the memory includes instructions executable by the processor to receive a reference signal ("RS") set for channel measurement corresponding to at least one non-zero power ("NZP") CSI-RS resource based on a CSI report setting; generate a CSI report based on the channel measurement, the CSI report including precoding matrix information for a precoding matrix and a set of coefficients corresponding to at least one dimension of the precoding matrix; and transmit the CSI report.
[0005] In one embodiment, a first method receives a set of RSs for channel measurement corresponding to NZP CSI-RS resources based on a CSI report setting, and generates a CSI report based on the channel measurement, wherein the CSI report includes precoding matrix information for a precoding matrix and a set of coefficients corresponding to at least one dimension of the precoding matrix. In one embodiment, the first method transmits the CSI report.
[0006] In one embodiment, the second device includes a processor and a memory coupled to the processor. In one embodiment, the memory includes instructions that can be executed by the processor to transmit to a UE a CSI report setting for configuring the UE for at least one of channel measurement and reporting; transmit to the UE an RS set for channel measurement corresponding to an NZP CSI-RS resource; and receive a CSI report from the UE. In one embodiment, the CSI report includes precoding matrix information for a precoding matrix and a coefficient set corresponding to at least one dimension of the precoding matrix, the precoding matrix including a first basis transform associated with a first dimension of the at least one dimension of the precoding matrix, the coefficient set corresponding to the precoding matrix decomposed into two coefficient subsets, the two coefficient subsets corresponding to different indices of the first dimension of the precoding matrix, a first subset of the two coefficient subsets associated with a second basis transform is associated with a second dimension of the precoding matrix, and a second subset of the two coefficient subsets associated with a third basis transform is associated with the second dimension of the precoding matrix.
[0007] In one embodiment, the second method transmits to the UE a CSI report setting for configuring the UE for at least one of channel measurement and reporting. In one embodiment, the second method transmits to the UE an RS set for channel measurement corresponding to an NZP CSI-RS resource. In one embodiment, the second method receives a CSI report from the UE, the CSI report comprising precoding matrix information for a precoding matrix and a coefficient set corresponding to at least one dimension of the precoding matrix, the precoding matrix comprising a first basis transform associated with a first dimension of the at least one dimension of the precoding matrix, the coefficient set corresponding to the precoding matrix decomposed into two coefficient subsets, the two coefficient subsets corresponding to different indices of the first dimension of the precoding matrix, a first subset of the two coefficient subsets associated with a second basis transform associated with a second dimension of the precoding matrix, and a second subset of the two coefficient subsets associated with a third basis transform associated with the second dimension of the precoding matrix. BRIEF DESCRIPTION OF THE DRAWINGS
[0008] Figure 1Illustrated is an example of a wireless communication system that supports techniques for joint CSI training and PMI feedback for AI-enabled networks in accordance with aspects of the present disclosure.
[0009] Figure 2 Illustrated is an example of aperiodic triggering states defining a CSI reporting settings list that supports techniques for joint CSI training and PMI feedback for AI-enabled networks in accordance with aspects of the present disclosure.
[0010] Figure 3 Illustrated is an example of aperiodic trigger state definition of resource sets and quasi-co-located (“QCL”) information indicative of techniques supporting joint CSI training and PMI feedback for AI-enabled networks in accordance with aspects of the present disclosure.
[0011] Figure 4 Illustrated is an example of a radio resource control (“RRC”) configuration of NZP CSI-RS resources that supports techniques for joint CSI training and PMI feedback for AI-enabled networks in accordance with aspects of the present disclosure.
[0012] Figure 5 Illustrated is an example of an RRC configuration of CSI interference measurement (“CSI-IM”) resources that supports techniques for joint CSI training and PMI feedback for AI-enabled networks in accordance with aspects of the present disclosure.
[0013] Figure 6 Illustrated is an example of partial CSI missing of Physical Uplink Shared Channel (“PUSCH”) based CSI that supports techniques for joint CSI training and PMI feedback for AI-enabled networks in accordance with aspects of the present disclosure.
[0014] Figure 7 Illustrated is an example of an NR protocol stack supporting techniques for joint CSI training and PMI feedback for AI-enabled networks in accordance with aspects of the present disclosure.
[0015] Figure 8 Illustrated is an example of a UE device supporting techniques for joint CSI training and PMI feedback for AI-enabled networks in accordance with aspects of the present disclosure.
[0016] Fig. 9 Illustrated is an example of a network equipment (“NE”) device that supports techniques for joint CSI training and PMI feedback for AI-enabled networks in accordance with aspects of the present disclosure.
[0017] Fig.10 A flow chart illustrating a method of supporting techniques for joint CSI training and PMI feedback for AI-enabled networks according to aspects of the present disclosure.
[0018] Fig.11 A flow chart illustrating a method of supporting techniques for joint CSI training and PMI feedback for AI-enabled networks according to aspects of the present disclosure. DETAILED DESCRIPTION
[0019] In general, the present disclosure describes systems, methods, and apparatus for techniques of joint CSI training and PMI feedback for AI-enabled networks. In a 3rd Generation Partnership Project ("3GPP") New Radio ("NR") network, CSI feedback in a frequency division dual ("FDD") network is reported by a UE to the network, where the CSI feedback is compressed via transforming the channel in the spatial domain, the frequency domain, or both using a predetermined set of spatial basis vectors and a predetermined set of frequency basis vectors, respectively. In addition to conventional CSI feedback mechanisms, AI / machine learning ("ML") enabled CSI acquisition schemes are considered as strong candidates for future generations of 3GPP NR networks.
[0020] Although AI / ML enabled schemes are often considered competitors to conventional CSI feedback schemes, most AI / ML enabled CSI acquisition schemes will still require some feedback from the UE to the network corresponding to CSI components that cannot be inferred from the AI / ML model (e.g., CSI components that are statistically independent over time and therefore cannot be inferred from the training data.) In addition, obtaining ubiquitous training data for the AI / ML enabled schemes is critical to maintaining the robustness of the AI / ML enabled CSI acquisition schemes to cope with changes in the environment that will cause shifts in channel allocations and therefore require updating the AI / ML model.
[0021] In view of the foregoing, a CSI feedback mechanism is proposed that aims to jointly provide CSI feedback from the network to the UE corresponding to the precoder codebook and provide training data to the network. Note that since the CSI feedback reported via the conventional method is compressed in the spatial domain, the frequency domain, or both, the CSI reported via the conventional CSI feedback scheme may not be as useful as the original CSI feedback.
[0022] In conventional solutions, CSI feedback (e.g., Type I codebook and Type II codebook) is used to train the AI / ML model. A disadvantage of this approach is that the CSI feedback in Type I codebook, Type II codebook is already compressed in the spatial domain, frequency domain, or both. Providing the model with CSI feedback compressed using a pre-configured compression method will limit the ability of the AI / ML model to design an appropriate compression scheme for the CSI feedback.
[0023] In another conventional solution, a new framework is specified for AI / ML-based CSI reporting configuration (e.g., specifying AI-based CSI configuration, measurement, and reporting). The disadvantage of this approach is the significant regulatory impact corresponding to defining a new measurement, reporting, and feedback framework.
[0024] The solution described in the present disclosure includes configuring a UE to provide feedback in the form of a CSI report, the CSI report including CSI corresponding to a codebook of a precoder, wherein a first CSI subset is reported in a compressed / transformed domain corresponding to a reduced CSI feedback overhead, and wherein a second CSI subset is reported in an uncompressed / untransformed domain such that the second CSI subset can be used for AI / ML model training in addition to codebook identification. Note that the two CSI subsets jointly correspond to the PMI, so both CSI subsets can still be used for precoder information feedback.
[0025] In one embodiment, based on AI / ML model inference at the network side, the network configures the UE with a subset of CSI compression types selected from a set of preconfigured CSI compression types, wherein the CSI compression types correspond to dimensions of the PMI, such as spatial dimensions, frequency / delay domain dimensions, and time / Doppler domain dimensions. Additionally, different quantization schemes for precoder coefficients may be configured based on AI / ML-based inference of the allocation of both amplitude and phase of the precoder coefficients.
[0026] Figure 1 A wireless communication system 100 is depicted that supports techniques for joint CSI training and PMI feedback for AI-enabled networks in accordance with an embodiment of the present disclosure. In one embodiment, the wireless communication system 100 includes at least one remote unit 105, a radio access network ("RAN") 120 (e.g., NG-RAN), and a mobile core network 140. The RAN 120 and the mobile core network 140 form a mobile communication network. The RAN 120 may be comprised of a base unit 110 with which the remote unit 105 communicates using a wireless communication link 115. Although in Figure 1 Although a specific number of remote units 105, base units 110, wireless communication links 115, RAN 120, and mobile core networks 140 are depicted in the figure, technicians in the field will recognize that any number of remote units 105, base units 110, wireless communication links 115, RAN 120, and mobile core networks 140 may be included in the wireless communication system 100.
[0027] In one embodiment, the RAN 120 is compatible with a 5G system specified in the 3GPP specification. In another embodiment, the RAN 120 is compatible with an LTE system specified in the 3GPP specification. However, more generally, the wireless communication system 100 may implement some other open or proprietary communication network (e.g., WiMAX) as well as other networks. The present disclosure is not intended to be limited to implementations of any particular wireless communication system architecture or protocol.
[0028] In one embodiment, the remote unit 105 may include a computing device, such as a desktop computer, a laptop computer, a personal digital assistant ("PDA"), a tablet computer, a smart phone, a smart TV (e.g., a TV connected to the Internet), a smart appliance (e.g., an appliance connected to the Internet), a set-top box, a game console, a security system (including a security camera), a car computer, a network device (e.g., a router, a switch, a modem), etc. In some embodiments, the remote unit 105 includes a wearable device, such as a smart watch, a fitness band, an optical head-mounted display, etc. In addition, the remote unit 105 may be referred to as a UE, a subscriber unit, a mobile device, a mobile station, a user, a terminal, a mobile terminal, a fixed terminal, a subscriber station, a user terminal, a wireless transmit / receive unit ("WTRU"), a device, or other terms used in the present technology.
[0029] The remote unit 105 may communicate directly with one or more of the base units 110 in the RAN 120 via uplink ("UL") communication signals and downlink ("DL") communication signals. In addition, the UL communication signals and the DL communication signals may be carried via wireless communication links 115. Here, the RAN 120 is an intermediate network that provides the remote unit 105 with access to the mobile core network 140.
[0030] In some embodiments, the remote unit 105 communicates with the application server 151 via a network connection with the mobile core network 140. For example, an application 107 in the remote unit 105 (e.g., a web browser, a media client, a phone / VoIP application) may trigger the remote unit 105 to create a PDU session (or other data connection) with the mobile core network 140 via the RAN 120. The mobile core network 140 then relays traffic between the remote unit 105 and the application server 151 in the packet data network 150 using the PDU session. Note that the remote unit 105 may create one or more PDU sessions (or other data connections) with the mobile core network 140. In this way, the remote unit 105 may simultaneously have at least one PDU session for communicating with the packet data network 150 and at least one PDU session for communicating with another data network (not shown).
[0031] The base units 110 may be distributed across a geographic region. In a particular embodiment, the base units 110 may also be referred to as access terminals, access points, base stations, base stations, node Bs, eNBs, gNBs, master node Bs, relay nodes, RAN nodes, or other terms used in the present technology. Generally speaking, the base units 110 are part of a RAN (e.g., RAN 120), which may include one or more controllers communicatively coupled to one or more corresponding base units 110. These and other elements of the radio access network are not illustrated, but the elements are generally well known to those skilled in the art. The base units 110 are connected to the mobile core network 140 via the RAN 120.
[0032] The base unit 110 may serve a number of remote units 105 within a service area (e.g., a cell or cell partition) via a wireless communication link 115. The base unit 110 may communicate directly with one or more of the remote units 105 via a communication signal. In general, the base unit 110 transmits DL communication signals to serve the remote units 105 in the time domain, frequency domain, and / or spatial domain. In addition, the DL communication signals may be carried via the wireless communication link 115. The wireless communication link 115 may be any suitable carrier in a licensed or unlicensed radio spectrum. The wireless communication link 115 facilitates communication between one or more of the remote units 105 and / or one or more of the base units 110. Note that during NR-U operation, the base unit 110 communicates with the remote units 105 via an unlicensed radio spectrum.
[0033] In one embodiment, the mobile core network 140 is a 5G core ("5GC") or an evolved packet core ("EPC"), which may be coupled to a packet data network 150, such as the Internet and personal data networks, among other data networks. The remote unit 105 may have a subscription or other account with the mobile core network 140. Each mobile core network 140 belongs to a single public land mobile network ("PLMN"). The present disclosure is not intended to be limited to implementation of any particular wireless communication system architecture or protocol.
[0034] The mobile core network 140 includes several network functions ("NFs"). As depicted, the mobile core network 140 includes a plurality of user plane functions ("UPFs") 141. The mobile core network 140 also includes a plurality of control plane functions including, but not limited to, an access and mobility management function ("AMF") 143 serving the RAN 120, a session management function ("SMF") 145, a policy control function ("PCF") 147, and a unified data management function ("UDM") 149.
[0035] In various embodiments, the mobile core network 140 supports different types of mobile data connections and different types of network slices, where each mobile data connection utilizes a specific network slice. As used herein, a "network slice" refers to a portion of the mobile core network 140 that is optimized for a specific traffic type or communication service. A network instance may be identified by an S-NSSAI, and a set of network slices that a remote unit 105 is authorized to use is identified by an NSSAI. In certain embodiments, various network slices may include separate instances of network functions (e.g., SMF 145 and UPF 141). In some embodiments, different network slices may share some common network functions, such as AMF 143. For ease of illustration, the network slices are described in detail in the figure. Figure 1 Different network slices are not shown in , but support for different network slices is assumed.
[0036] Despite Figure 1 Although a specific number and type of network functions are depicted in the mobile core network 140, those skilled in the art will recognize that any number and type of network functions may be included in the mobile core network 140. Furthermore, when the mobile core network 140 is an EPC, the depicted network functions may be replaced with appropriate EPC entities (e.g., MME, S-GW, P-GW, HSS, etc.). In a particular embodiment, the mobile core network 140 may include an AAA server.
[0037] Although Figure 1 Components of a 5G RAN and 5G core network are depicted, but the described embodiments are applicable to other types of communication networks and RATs, including IEEE 802.11 variants, GSM, GPRS, UMTS, LTE variants, CDMA 2000, Bluetooth, Zigbee, Sigfoxx, etc. For example, in an LTE variant involving EPC, the AMF may be mapped to the MME, the SMF to the control plane portion of the PGW and / or to the MME, the UPF to the SGW and the user plane portion of the PGW, the UDM / UDR to the HSS, etc.
[0038] In the following description, the term "gNB" is used as a base station, but can be replaced by any other radio access node such as RAN node, eNB, BS, eNB, gNB, AP, NR, etc. In addition, the operation is mainly described in the context of 5G NR. However, the proposed solution / method is also equally applicable to other mobile communication systems that support CSI reporting.
[0039] Regarding the NR Rel.15 Type II codebook, assume that the gNB is equipped with a 2D antenna array having N1 and N2 antenna ports in each polarization direction placed horizontally and vertically, and communication occurs on N3 PMI sub-bands. A PMI sub-band consists of a set of resource blocks, and each resource block consists of a set of subcarriers. In this case, for the NR Rel.15 Type II codebook, 2N1N2 CSI-RS ports are used to achieve high-resolution DL channel estimation. To reduce UL feedback overhead, DFT-based CSI compression in the spatial domain is applied to L dimensions in each polarization direction, where L < N1N2. Hereinafter, the indices of the 2L dimensions are referred to as SD basis indices. The magnitude and phase values of the linear combination coefficients for each sub-band are fed back to the gNB as part of the CSI report. The 2N1N2xN3 codebook for each layer is in the following form:
[0040] W = W1W2,
[0041] where W1 is a 2N1N2x2L block diagonal matrix with two identical diagonal blocks (L < N1N2), i.e.,
[0042]
[0043] and B is an N1N2xL matrix whose columns are taken from a 2D oversampled DFT matrix, as follows.
[0044]
[0045]
[0046]
[0047]
[0048]
[0049] where the superscript T denotes the matrix transpose operation. Note that for the 2D DFT matrix from which the matrix B is extracted, assume oversampling factors O1 and O2. Note that W1 is common across all layers. W2 is a 2LxN3 matrix, where the i-th column corresponds to the linear combination coefficients of the 2L beams in the i-th sub-band. Only the indices of the L selected columns of B, and the oversampling index taking the value of O1O2, are reported. Note that W2 is independent for different layers.
[0050] Regarding the NR Rel.15 Type II port selection codebook, only K (where K≤2N1N2) beamformed CSI-RS ports are utilized in DL transmission to reduce complexity. Each layer of the KxN3 codebook matrix is in the following form:
[0051]
[0052] Here, W2 follows the same structure as the conventional NR Rel.15 Type II codebook and is layer specific. is a Kx2L block diagonal matrix with two identical diagonal blocks, i.e.,
[0053]
[0054] And E is The matrix, whose columns are standard unit vectors, is as follows.
[0055]
[0056] in is a standard unit vector with 1 at the ith position. Here, d PS is in PS ≤min(K / 2, L) under the condition that the RRC parameters take the values {1, 2, 3, 4}, and m PS Value and is reported as part of the UL CSI feedback overhead. W1 is common across all layers.
[0057] For K=16, L=4 and d PS =1, corresponding to m PS = {0,1,…,7}, the 8 possible implementations of E are as follows
[0058]
[0059] When PS =2, corresponding to m PS = {0,1,2,3}, the four possible implementations of E are as follows
[0060]
[0061] When PS =3, corresponding to m PS = {0,1,2}, the three possible implementations of E are as follows
[0062]
[0063] When PS =4, corresponding to m PSTwo possible implementations of E = {0,1} are as follows
[0064]
[0065] Generally speaking, m PS parameterizes the position of the first 1 in the first column of E, while d PS represents the row shift corresponding to different values of m PS of the corresponding rows
[0066] The NR Rel.15 Type I codebook is the baseline codebook for NR with various configurations. The most common utility of the Rel.15 Type I codebook is a special case of the NR Rel.15 Type II codebook, where, for RI = 1, 2, L = 1, where the phase coupling values are reported for each subband, i.e., W2 is 2xN3, where the first column is equal to [1, 1, …, 1] and the second row is equal to Under specific configurations, φ0 = φ1 … = φ, i.e., wideband reporting. For RI > 2, different beams are used for each pair of layers. Obviously, the NR Rel.15 Type I codebook can be depicted as a low-resolution version of the NR Rel.15 Type II codebook, where only the spatial beam selection and phase combination for each layer pair are involved
[0067] Regarding the NR Rel.16 Type II codebook, it is assumed that the gNB is equipped with a two-dimensional ("2D") antenna array having N1, N2 antenna ports in each polarization direction placed horizontally and vertically and communication occurs on N3 PMI subbands. The PMI subbands consist of sets of resource blocks, and each resource block consists of a set of subcarriers. In this case, for the NR Rel.16 Type II codebook, 2N1N2N3 CSI-RS ports are used to achieve high-resolution DL channel estimation. To reduce the UL feedback overhead, DFT-based CSI compression in the spatial domain is applied to L dimensions in each polarization direction, where L < N1N2. Similarly, additional compression in the frequency domain is applied, where each beam of the frequency-domain precoding vector is transformed into the delay domain using the inverse DFT matrix, and a subset of the magnitude and phase values of the delay-domain coefficients is selected and fed back to the gNB as part of the CSI report. The 2N1N2xN3 codebook for each layer is in the following form:
[0068]
[0069] where W1 is a 2N1N2x2L block diagonal matrix with two identical diagonal blocks (L < N1N2), i.e.,
[0070]
[0071] And B is an N1N2xL matrix whose columns are taken from a 2D oversampled DFT matrix, as follows.
[0072]
[0073]
[0074]
[0075]
[0076]
[0077] where the superscript T denotes the matrix transpose operation. Note that for the 2D DFT matrix from which matrix B is extracted, assume oversampling factors O1, O2. Note that W1 is shared across all layers.
[0078] W f is an N3xM matrix (M < N3) whose columns are selected from a critically sampled size N3 DFT matrix, as follows:
[0079]
[0080]
[0081] Only report the indices of the L selected columns of B, and the oversampling indices taking the O1O2 values. Similarly, for W F , only report the indices of the M columns selected from a DFT matrix of predefined size N3. In the following text, the indices of the M dimensions are referred to as the selected frequency domain ("FD") basis indices. Thus, L and M represent the equivalent spatial dimension and frequency dimension after compression, respectively. Finally, the 2LxM matrix represents the linear combination coefficients ("LCC") of the spatial and frequency DFT basis vectors. W f is selected independently for different layers. The magnitude and phase values of an approximate β fraction of the 2LM available coefficients are reported to the gNB as part of the CSI report (β < 1). Note that the coefficients with zero magnitude are indicated via a per-layer bitmap. Since all the coefficients reported within a layer are normalized with respect to the coefficient with the largest magnitude (the strongest coefficient), the relative values of the coefficients are set to 1, and for this coefficient, no magnitude or phase information is explicitly reported. Only the indication of the index of the strongest coefficient for each layer is reported. Thus, for single-layer transmission, each layer reports the magnitude and phase values of the maximum of the coefficients (and the indices of the selected L and M DFT vectors), resulting in a significant reduction in the CSI report size compared to reporting the information of 2N1N2xN3 - 1 coefficients.
[0082] Regarding the NR Rel.16 Type II port selection codebook, only K (K≤2N1N2) beamformed CSI-RS ports are utilized in DL transmission to reduce complexity. Each layer of the KxN3 codebook matrix is in the following form:
[0083]
[0084] Here, and W3 follow the same structure as the conventional NR Rel.16 Type II codebook, where both are layer specific. It is a Kx2L block diagonal matrix with the same structure as the Kx2L block diagonal matrix in the NR Rel.15 Type II port selection codebook.
[0085] The codebook report is split into two parts based on the priority of the reported information. Each part is coded separately (part 1 has a possibly higher code rate). Only the parameters for NR Rel.16 Type II codebook are listed below. More details can be found in TS 38.214 Sec 5.2.3-4.
[0086] The CSI report contains:
[0087] a. Part 1: RI+CQI+total number of coefficients
[0088] b. Part 2: SD base indicator + FD base indicator / layer + bitmap / layer + coefficient amplitude information / layer + coefficient phase information / layer + strongest coefficient indicator / layer
[0089] Furthermore, Part 2 CSI may be decomposed into several sub-parts each with different priorities (higher priority information is listed first). This segmentation is required to allow for dynamic reporting size of codebooks based on available resources in the uplink phase. More details can be found in TS 38.214 Sec 5.2.3.
[0090] In addition, the type II codebook is based on aperiodic CSI reporting and reports in PUSCH only via DCI triggering (one exception). The type I codebook can be based on periodic CSI reporting (e.g., physical uplink control channel ("PUCCH")) or semi-persistent CSI reporting (e.g., PUSCH or PUCCH) or aperiodic reporting (e.g., PUSCH).
[0091] Regarding priority reporting of Part 2 CSI, note that multiple CSI reports may be transmitted, as shown in Table 1:
[0092]
[0093]
[0094] Table 1: CSI reporting priority order
[0095] Note that N Rep The priority of CSI reports is based on: a CSI report corresponding to one CSI reporting configuration for one cell may have a higher priority than another CSI report corresponding to another CSI reporting configuration for the same cell; a CSI report intended for one cell may have a higher priority than another CSI report intended for another cell; a CSI report may have a higher priority based on the CSI report content, for example, the CSI report carries L1-RSRP information with a higher priority; and a CSI report may have a higher priority based on its type, for example, whether the CSI report is non-periodic, semi-persistent or periodic and whether the report is sent via PUSCH or PUCCH may affect the priority of the CSI report.
[0096] In view of this, CSI reports can be prioritized as follows, where CSI reports with lower IDs have higher priority:
[0097] Pri iCSI (y, k, c, s) = 2·N cells ·M s y+N cells ·M s ·k+M s ·c+s
[0098] s: CSI report configuration index, and M s : Maximum number of CSI report configurations
[0099] c: cell index, and N cells : Number of service cells
[0100] k: 0 for CSI reporting carrying L1-RSRP or L1-SINR, 1 for other cases
[0101] y: 0 for aperiodic reporting, 1 for semi-persistent reporting on PUSCH, 2 for semi-persistent reporting on PUCCH, and 3 for periodic reporting.
[0102] Regarding triggered aperiodic CSI reporting on PUSCH, the UE needs to use the CSI framework in NR Rel.15 to report the required CSI information to the network. The triggering mechanism between the reporting setting and the resource setting can be summarized in Table 2 below.
[0103]
[0104] Table 2: Trigger mechanism between report settings and resource settings
[0105] In addition, in one embodiment, all associated resource settings of a CSI reporting setting need to have the same time domain behavior. Once configured by RRC, periodic CSI-RS / IM resources and CSI reporting are assumed to be present and active, aperiodic and semi-persistent CSI-RS / IM resources and CSI reporting should be explicitly triggered or activated, aperiodic CSI-RS / IM resources and aperiodic CSI reporting are jointly triggered via transmission DCI format 0-1, and semi-persistent CSI-RS / IM resources and semi-persistent CSI reporting are independently activated.
[0106] Figure 2 1 illustrates an example of an aperiodic trigger state that defines a CSI report setting list that supports techniques for joint CSI training and PMI feedback for AI-enabled networks in accordance with aspects of the present disclosure. Specifically, diagram 200 includes DCI format 0_1 202, CSI request code point 204, and aperiodic trigger state 2 206. In addition, aperiodic trigger state 2 includes ReportConfigID x 208, ReportConfigID y 210, and ReportConfigID z 212. For aperiodic CSI-RS / IM resources and aperiodic CSI reports, triggering is accomplished jointly by transmitting DCI formats 0-1. DCI format 0_1 contains a CSI request field (0 to 6 bits). A non-zero request field points to a so-called aperiodic trigger state configured by RRC. The aperiodic trigger state is then defined as a list of up to 16 aperiodic CSI reporting configurations, identified by a CSI reporting configuration ID, for which the UE simultaneously calculates the CSI and transmits the CSI on the scheduled PUSCH transmission.
[0107] Figure 3 An example of a non-periodic trigger state definition of resource sets and QCL information indicating techniques for supporting joint CSI training and PMI feedback for AI-enabled networks according to aspects of the present disclosure is illustrated. In particular, Figure 3Depicts an aperiodic trigger state indicating a resource set 302 and QCL information 304. When a CSI reporting setting is linked with an aperiodic resource setting (which may include multiple resource sets), the aperiodic NZP-CSI-RS resource set for channel measurement, the aperiodic CSI-IM resource set (if used), and the aperiodic NZP CSI-RS resource set for IM (if used) to be used for a given CSI reporting setting are also included in the aperiodic trigger state definition. For the aperiodic NZP-CSI-RS, the QCL source to be used is also configured in the aperiodic trigger state. The UE assumes that the resources used to calculate the channel and interference can be processed with the same (i.e., quasi-co-located with respect to "QCL-TypeD") spatial filter.
[0108] Figure 4 An example of an RRC configuration of NZP CSI-RS resources that supports techniques for joint CSI training and PMI feedback for AI-enabled networks in accordance with aspects of the present disclosure is illustrated. In particular, Figure 4 A code sample 402 is depicted illustrating an embodiment of a process of aperiodic triggering status indicating a set of resources and QCL information.
[0109] Figure 5 An example of an RRC configuration of CSI-IM resources that supports techniques for joint CSI training and PMI feedback for AI-enabled networks in accordance with aspects of the present disclosure is illustrated. In particular, Figure 4 A code sample 502 is depicted illustrating an embodiment of an RRC configuration including NZP-CSI-RS resources and CSI-IM-resources.
[0110] The types of uplink channels used for CSI reporting as a function of the CSI codebook type are summarized in Table 3 below:
[0111]
[0112] Table 3: Uplink channels used for CSI reporting as a function of CSI codebook type
[0113] For aperiodic CSI reporting, PUSCH-based reporting is divided into two CSI parts: CSI Part 1 and CSI Part 2. The reason for this is that the size of the CSI payload varies significantly, and thus the worst-case UCI payload size design will result in a large overhead.
[0114] CSI Part 1 has a fixed payload size (and can be decoded by the gNB without a priori information) and contains the RI (if reported), CRI (if reported) and CQI for the first codeword, and a certain number of non-zero wideband amplitude coefficients for each layer for Type IICSI feedback on PUSCH.
[0115] CSI part 2 has a variable payload size that can be derived from the CSI parameters in CSI part 1 and contains the PMI and CQI for the second codeword when RI>4.
[0116] For example, if the aperiodic trigger state indicated by DCI format 0_1 defines 3 reporting settings x, y and z, then the aperiodic CSI report for CSI part 2 will be as follows: Figure 6 Determine the order as indicated in.
[0117] Figure 6 An example of partial CSI omission of PUSCH-based CSI that supports techniques for joint CSI training and PMI feedback for AI-enabled networks in accordance with aspects of the present disclosure is illustrated. In particular, Figure 6 6 is a schematic block diagram 600 illustrating an embodiment of partial CSI omission for PUSCH based CSI. The diagram 600 includes ReportConfigIDx 602, ReportConfigIDy 604, and ReportConfigIDz 606. Furthermore, the diagram 600 includes a first report 608 (e.g., a requested quantity to be reported) corresponding to ReportConfigIDx 602, a second report 610 (e.g., a requested quantity to be reported) corresponding to ReportConfigIDy 604, and a third report 612 (e.g., a requested quantity to be reported) corresponding to ReportConfigIDz 606. Each of the first report 608, the second report 610, and the third report 612 includes CSI part 1 620 and CSI part 2 622. The order 623 of CSI part 2 across the number of reports is CSI part 2 of the first report 624, CSI part 2 of the second report 626, and CSI part 2 of the third report 628. In addition, the CSI part 2 report can generate Report 1 WB CSI 634, Report 2 WB CSI 636, Report 3 WB CSI 638, Report 1 Even SB CSI 640, Report 1 Odd SB CSI 642, Report 2 Even SB CSI 644, Report 2 Odd SB CSI 646, Report 3 Even SB CSI 648 and Report 3 Odd SB CSI 650.
[0118] As mentioned above, CSI reports are prioritized based on: (1) time domain behavior and physical channel, where more dynamic reports take precedence over less dynamic reports and PUSCH takes precedence over PUCCH; (2) CSI content, where beam reports (i.e., L1-RSRP reports) take precedence over regular CSI reports; (3) the serving cell to which the CSI corresponds (in the case of CA operation). CSI corresponding to PCell takes precedence over CSI corresponding to Scell; and (4) CSI report configuration IDreportConfigID.
[0119] In the present disclosure, the embodiments described below are directed to a CSI feedback mechanism that aims to jointly provide CSI feedback from the network to a UE corresponding to a precoder codebook and provide training data to the network. Note that since the CSI feedback reported via conventional methods is compressed in the spatial domain, the frequency domain, or both, the CSI reported via the conventional CSI feedback scheme may not be as useful as the original CSI feedback.
[0120] More specifically, the subject matter herein is directed to configuring a UE to feed back a CSI report, wherein the CSI report includes CSI corresponding to a code book of a precoder, wherein a first CSI subset is reported in a compressed / transformed domain for reducing CSI feedback overhead, and wherein a second CSI subset is reported in an uncompressed / untransformed domain, such that the second CSI subset can be used for AI / ML model training in addition to code book identification; and based on AI / ML model inference at the network side, the network configures a CSI compression type subset selected from a preconfigured CSI compression type set for the UE, wherein the CSI compression type corresponds to dimensions of the PMI, for example, spatial dimensions, frequency / delay domain dimensions, and time / Doppler domain dimensions.
[0121] Several embodiments and examples are provided to explain the proposal and to illustrate how the proposal can be employed in practical situations.
[0122] In general, given CSI feedback for a channel with 32 ports and 13 SBs, the first step would be to transform the channel using spatial DFT → new dimension is 8x13. Step 2 is to feed back 13 coefficients corresponding to the strongest beam out of the 8 beams. Step 3 is to further transform 7 beams using delay domain DFT → new dimension is 7x4. Step 4 is to feed back 28 coefficients corresponding to the weakest 7 beams.
[0123] In one embodiment, it is assumed that the channel between the UE and the gNB has an occupation N SB Frequency band (index n=0, ..., N SB-1) with P channel paths (index p=0, ..., P-1), where the gNB is equipped with K antennas (index k=0, ..., K-1). Then, the channel at time index δ can be expressed as follows:
[0124]
[0125] where g k,p is the complex gain of path p at antenna k; Δf is the PMI subband spacing; τ p is the delay of path p; F c is the carrier frequency; c is the speed of light; d is the antenna spacing at the gNB; θ p is the angular spatial displacement of the gNB antenna array corresponding to path p; δ is the time index; v is the relative velocity between the gNB and the UE; and Φ p It is the angle between the moving direction of path p and the incident direction of the signal.
[0126] The above channels are parameterized by the following three dimensions: spatial dimension, frequency dimension and time dimension. In order to construct a precoder codebook with reasonable CSI feedback overhead, the CSI corresponding to the three dimensions needs to be compressed. In the Rel.16 type II codebook, both the spatial domain and the frequency domain are compressed via DFT transforms of the spatial domain and the frequency domain having columns of a two-dimensional DFT matrix and a one-dimensional DFT matrix, respectively. However, since the impact of the time variation of the channel is negligible at low UE speeds (e.g., small v), the time variation of the channel is often ignored in the codebook design. This approximation may not be efficient at high UE speeds or due to rapid changes in UE orientation. Similarly, in an environment with fewer scatterers and higher UE speeds, reporting CSI about multiple spatial and time / Doppler domain indices may be more important than reporting frequency / delay domain indices, because channels with fewer scatterers are more likely to cause frequency flat fading, for example, the channel coefficients corresponding to a common spatial domain index and a common time domain index are strongly correlated.
[0127] In the present disclosure, a framework is proposed to implement joint CSI and training data feedback, wherein the training data is inherent in the CSI feedback, and wherein the training data enables the network to select one or more CSI feedback compression scheme subsets for one or more of the spatial domain dimensions, frequency / delay domain dimensions, or time / Doppler domain dimensions from a set of preconfigured CSI feedback schemes. The following describes several embodiments of a CSI feedback framework that provides joint CSI and training data feedback and further enables the network to select CSI feedback compression schemes for one or more spatial domain dimensions, frequency / delay domain dimensions, or time / Doppler domain dimensions from a set of preconfigured CSI feedback schemes. According to a possible embodiment, one or more elements or features from one or more of the described embodiments may be combined.
[0128] In the first embodiment of the AI / ML based CSI framework, there are multiple alternatives to outline the AI / ML algorithm functionality. In one embodiment, the AI / ML model is trained at the UE node. Since the UE is a node that can seamlessly collect training data for CSI acquisition using DL pilot signals (e.g., CSI-RS for channel measurement), this alternative may seem reasonable, however, since new AI / ML models should be retrained whenever the environment changes (e.g., change in UE position / orientation), the memory and computational complexity requirements for this operation will be huge.
[0129] In another embodiment, the AI / ML model is trained at the network node. One advantage of this approach is that the network has significantly more power and computing capabilities than UE nodes, and can therefore manage training moderately complex AI / ML models, and storing large amounts of training data. In addition, since the network nodes are mostly assumed to be fixed, their coverage area is expected to be the same, and therefore the same AI / ML model can be applied to several UEs within a specific area of the cell within a reasonable period of time. One challenge with this approach involves obtaining training data at the network nodes, especially for FDD systems where UL / DL channel reciprocity cannot be maintained. Note that when evaluating the efficiency of the AI / ML algorithm, the overhead corresponding to feeding back the training data from the UE to the network should be considered as one of the metrics.
[0130] In the following, due to the advantages corresponding to the storage, computation and cell-centric nature of network-based AI / ML model computation, it is assumed that the AI / ML model is trained at the network. The challenges corresponding to obtaining training data corresponding to the DL channel at the network side are discussed in the next section.
[0131] Assuming that the AI / ML model is trained at the network, several aspects of DL training data acquisition at the network side are discussed to achieve efficient AI / ML modeling. In one embodiment, to maintain the robustness of the AI / ML model with respect to channel changes, DL training data should be continuously fed back to the network to keep up with changes in the environment (e.g., traffic, weather, and mobile scatterers). Note that this may not necessarily correspond to online learning; even for offline learning algorithms, a framework for obtaining new training data corresponding to channel changes should be characterized.
[0132] In another embodiment, based on the current codebook-based DL CSI feedback scheme in NR, the CSI is compressed in at least one or both of the spatial domain or the frequency domain. An intuitive approach would be to use codebook-based CSI feedback (e.g., type I codebook and / or type II codebook) to obtain training data. One disadvantage of this approach is that the training data will include CSI feedback that has been compressed by conventional methods, which will have a detrimental effect on the AI / ML model inference accuracy. For example, if the AI / ML model compares the output of the AI / ML model with the channel corresponding to the CSI feedback to evaluate the AI / ML model inference accuracy, then this evaluation will not be accurate because the evaluation is based on the channel estimate H' based on the predefined compression, rather than the digital quantized channel H in the spatial domain or the frequency domain without further compression. On the other hand, if the UE feeds back the training data corresponding to the DL CSI feedback without compression in the spatial dimension and / or the frequency dimension, then the feedback overhead of the training data will be significant, which will violate the purpose of using the AI / ML model primarily to reduce the overall CSI feedback overhead. Numerically, AI / ML-based CSI feedback aims to minimize the following metrics:
[0133]
[0134] Where H represents the digital domain representation of the channel matrix. On the other hand, the compressed channel H' representing the channel recovered after the codebook-based transform will produce the following optimization metric:
[0135]
[0136] Since H≠H′, the two optimized outputs will produce different channel estimates. In the following, a new method is proposed to balance the quality of training data and its corresponding feedback overhead.
[0137] In one embodiment, for DL CSI acquisition in NR, whether the network operates in FDD mode or time division duplex ("TDD") mode, AI / ML will not be able to completely replace RS-based CSI feedback for high-resolution precoding design, because some channel parameters may vary from one time instant to another, without strong correlation across two time instants (e.g., the initial random phase of the channel). In view of this, an AI / ML-based CSI framework can be envisioned as a means to further reduce CSI feedback overhead compared to conventional methods, for example, by reducing the number of dominant spatial domain basis indices, frequency / delay domain basis indices, and time / Doppler domain basis indices after spatial domain transforms, frequency domain transforms, and time domain transforms, respectively. Although the current CSI feedback framework already provides CSI feedback overhead reduction by utilizing such transforms, the CSI dimensionality can be further reduced if a wider range of transform techniques are preconfigured (where different transforms can be selected for a given UE based on channel variations).
[0138] In this section, a CSI feedback framework method is proposed that simultaneously enables codebook reporting and training data feedback from the UE to the network. Under this framework, the UE is configured with a CSI reporting configuration, which includes reporting a codebook corresponding to a PMI, where a subset of the PMI is used as CSI training data. Note that this PMI subset can be used for both precoder identification and training data. In order to make this PMI subset usable for both purposes, the report has different characteristics compared to the characteristics of the rest of the PMI. To elaborate on these characteristics in more detail, two alternative NR codebook designs are reviewed, in which the network node is equipped with a two-dimensional (2D) antenna array having N1, N2 antenna ports in each polarization direction placed horizontally and vertically, respectively, and communication occurs on N3 subbands.
[0139] In codebook design 1 involving spatial domain CSI compression, the first design of the PMI codebook for a given layer is as follows:
[0140] W=W1W2
[0141] Under this design, the codebook W includes a spatial domain transformation matrix W1 with size 2N1N2×2L, transforming the 2N1N2 dimensional spatial domain into a transformed spatial domain with 2L dimensions, where L≤N1N2, and a matrix W2 with size 2L×N3 includes the linear combination coefficients of the codebook, whose rows and columns correspond to the transformed spatial domain index and frequency domain index, respectively. Note that this design is similar to the Rel.15 type II codebook, which only compresses the spatial domain of the codebook.
[0142] In codebook design 2 involving frequency domain CSI compression, the second design of the PMI codebook for a given layer is as follows:
[0143]
[0144] Under this design, the codebook W consists of a frequency domain transform matrix W with size N3×M f , the frequency domain transformation matrix transforms the N3-dimensional frequency domain into a transformed frequency domain having M dimensions, where M≤N3, and has a matrix size of 2N1N2×M The matrix W includes linear combination coefficients of a codebook whose rows and columns correspond to spatial domain indices and transformed frequency domain indices, respectively. f It may correspond to a frequency domain transform, a time / Doppler domain transform, or a combination thereof.
[0145] In codebook design 3 involving joint spatial domain CSI compression and frequency domain CSI compression, the third design of the PMI codebook for a given layer is as follows:
[0146]
[0147] Under this design, the codebook W includes a spatial domain transform matrix W1 with size 2N1N2×2L, which applies a spatial domain transform similar to that of Design 1. In addition, a second transform matrix W with size N3×M is applied f The second change matrix applies a frequency domain transform similar to codebook design 2. A matrix with size 2L×M The matrix W comprises linear combination coefficients of a codebook whose rows and columns correspond to transformed spatial domain indices and transformed frequency domain indices, respectively. f It may correspond to a frequency domain transform, a time / Doppler domain transform, or a combination thereof.
[0148] Several embodiments describing channel transformation on at least one of the transformed spatial domain dimension, the transformed frequency / delay domain dimension and the transformed time / Doppler domain dimension are provided below. According to a possible embodiment, one or more elements or features from one or more of the described embodiments may be combined.
[0149] In a first embodiment, a set of non-zero channel coefficients corresponding to a PMI codebook is decomposed into two subsets of channel coefficients, wherein the coefficients corresponding to a first of the two subsets of channel coefficients are associated with a first set of indices of a first dimension of the PMI, and wherein a second of the two subsets of channel coefficients is associated with a second set of indices of the first dimension of the PMI.
[0150] In a first example, a first dimension of the PMI corresponds to a transformed spatial domain, where two subsets of channel coefficients are associated with different sets of index values corresponding to the transformed spatial domain index.
[0151] In a second example, the first dimension of the PMI corresponds to a transformed frequency / delay domain, wherein two subsets of channel coefficients are associated with different sets of index values corresponding to the transformed frequency / delay domain indices.
[0152] In a third example, the first dimension of the PMI corresponds to a transformed time / Doppler domain, where two subsets of channel coefficients are associated with different sets of index values corresponding to the transformed time / Doppler domain indices.
[0153] In a fourth example, the first dimension of the PMI corresponds to a joint domain corresponding to a transformed frequency / delay domain and a transformed time / Doppler domain, wherein two subsets of channel coefficients are associated with different sets of index values corresponding to the joint transformed frequency / delay domain and transformed time / Doppler domain indices.
[0154] In a fifth example, the first subset of channel coefficients corresponds to a first codebook design, e.g., codebook design 1, and the second subset of channel coefficients corresponds to a second codebook design, e.g., codebook design 2. Alternatively, the first subset of channel coefficients corresponds to codebook design 2, and the second subset of channel coefficients corresponds to codebook design 3.
[0155] In a second embodiment, the first set of indices for the first dimension of the PMI is disjoint with respect to the second set of indices for the first dimension of the PMI. In a first example, the first set of indices for the first dimension of the PMI includes a single index, e.g., an index corresponding to a maximum power or a channel gain, and the second set of indices for the first dimension of the PMI includes the remainder of the selected indices, i.e., all selected indices except the index corresponding to the first set of indices.
[0156] In a third embodiment, a first of the two subsets of non-zero channel coefficients corresponds to a first transform of the first dimension of the PMI, and a second of the two subsets of non-zero coefficients corresponds to a second transform of the first dimension of the PMI.
[0157] In a first example, the first transform is a trivial transform, eg, a scaling of coefficients corresponding to the first set of indices.
[0158] In a second example, the first transform corresponds to averaging of channel coefficients corresponding to the first set of indices.
[0159] In a third example, the first transform corresponds to selection of channel coefficients corresponding to a first set of indices, eg, selection of even indices, or selection of odd indices.
[0160] In a fourth example, the second transform corresponds to a DFT matrix transform.
[0161] In a fifth example, the second transform corresponds to a discrete wavelet transform ("DWT") matrix transform.
[0162] In a sixth example, the second transform corresponds to a discrete sine / cosine transform ("DST" / "DCT") matrix transform.
[0163] As discussed in the previous section, CSI feedback is classified into two parts based on whether the CSI feedback corresponds to training-based CSI feedback or PMI-based CSI feedback. In view of this, the training-based CSI feedback corresponding to the first of the two channel coefficient subsets can help the network select the best transform / compression method for the PMI-based CSI feedback corresponding to the second of the two channel coefficient subsets. Several embodiments are described below. According to a possible embodiment, one or more elements or features from one or more of the described embodiments may be combined.
[0164] In a first embodiment, the training-based CSI feedback includes a larger number of coefficients per dimension than the PMI-based CSI feedback.
[0165] In a second embodiment, coefficients corresponding to the training-based CSI feedback are quantized using scalar quantization, e.g., each of an amplitude value and a phase value corresponding to a given coefficient is encoded separately. In one example, the coefficients corresponding to the training-based CSI feedback are quantized with a higher resolution than the PMI-based CSI feedback, e.g., the coefficients corresponding to the training-based CSI feedback are quantized with a larger number of bits than the coefficients corresponding to the PMI-based CSI feedback.
[0166] In a third embodiment, coefficients corresponding to PMI-based CSI feedback are quantized using a codebook from a set of preconfigured codebooks. In one example, the set of preconfigured codebooks includes one of a codebook with amplitude values uniformly distributed in a linear domain and a codebook with amplitude values uniformly distributed in a logarithmic domain.
[0167] In a fourth embodiment, coefficients corresponding to PMI-based CSI feedback are quantized using vector quantization (eg, amplitude values corresponding to two coefficients, or phase values corresponding to two coefficients, or some combination thereof) and jointly encoded.
[0168] In a fifth embodiment, a UE configured with training-based CSI feedback is also expected to be configured with a CSI domain transform based on the training-based CSI feedback across one or more of the spatial domain, frequency domain, or time domain corresponding to the PMI-based CSI feedback. In a first example, a CSI report setting received at the UE includes configuration parameters that configure the UE with a transform scheme from a set of preconfigured transform schemes corresponding to at least one CSI dimension. In a second example, the set of preconfigured transform schemes includes one or more of a DFT scheme, a DCT, a DST, a DWT, a base averaging transform, a weighted averaging transform, or some combination thereof.
[0169] For a transformation matrix that transforms one dimension of a channel from N indices to N' indices (N'≤N), the following example is provided:
[0170] Example 1: One-dimensional DFT matrix transformation:
[0171]
[0172]
[0173] Example 2: One-dimensional DCT matrix transformation:
[0174]
[0175]
[0176] Example 3: One-dimensional DST matrix transformation:
[0177]
[0178]
[0179] Example 4: One-dimensional Haar matrix transformation:
[0180]
[0181]
[0182]
[0183] Among them I S corresponds to an identity matrix of size s, and the operator is the Kronecker product.
[0184] For a transform matrix that transforms two dimensions (eg, joint time / Doppler and frequency domains), the following example is provided:
[0185] Example 1: Two-dimensional DFT matrix transformation:
[0186]
[0187]
[0188]
[0189] Example 2: Two-dimensional DCT matrix transformation:
[0190]
[0191]
[0192]
[0193] Example 3: Two-dimensional DST matrix transformation:
[0194]
[0195]
[0196]
[0197] In a sixth embodiment, parameters corresponding to indices of a selected subset of basis vectors of the domain transform matrix are communicated between the UE and the network. In a first example, the parameters are fed back from the UE to the network as part of a CSI report. In a second example, the parameters are higher layer configurations from the network to the UE. In a third example, the parameters are represented as combined values, e.g., k basis vectors are selected from a set of N basis vectors, where k ≤ N, via Bit representation, where and and Corresponds to the smallest integer value greater than or equal to the variable a, e.g., an upper limit function.
[0198] Different embodiments are presented discussing the content and configuration of CSI reports. Note that combinations of elements of one or more embodiments are not excluded.
[0199] In a first embodiment, the CSI report corresponding to the hybrid-joint training data and PMI feedback is configured via a CSI report configuration having one or more of the following quantities: "CRI", "RI", "PMI", "CQI", "LI", "SSBRI", "L1-SINR", "L1-RSRP". One or more additional quantities (e.g., "TDI") corresponding to the training data indication may also be configured.
[0200] In a second embodiment, the CSI report corresponding to the hybrid combined training data and PMI feedback includes the location information of the UE. In the first example, the CSI feedback report corresponding to one codebook includes the differential location information relative to the location of the UE at the time of the last CSI feedback report corresponding to the a priori codebook transmission. In the second example, the report location information is based on a higher layer configuration within the CSI reporting configuration.
[0201] In a third embodiment, a CSI report corresponding to hybrid joint training data and PMI feedback includes a bitmap indicating indices of non-zero coefficients, and wherein coefficients corresponding to training-based CSI feedback are not included in the bitmap. In one example, coefficients corresponding to training-based CSI feedback are always reported.
[0202] In a fourth embodiment, a CSI reporting configuration corresponding to hybrid joint training data and PMI feedback includes enabling and disabling higher layer parameters including training-based CSI feedback corresponding to a first of two channel coefficient subsets.
[0203] In some embodiments, the terms antenna, panel, and antenna panel may be used interchangeably with respect to antenna panels / ports, quasi-colocation, TCI states, spatial relationships. An antenna panel may be hardware for transmitting and / or receiving radio signals at frequencies below 6 GHz (e.g., FR1) or above 6 GHz (e.g., FR2 or millimeter waves). In some embodiments, an antenna panel may include an array of antenna elements, each of which is connected to hardware, such as a phase shifter that allows a control module to apply spatial parameters to the transmission and / or reception of a signal. The resulting radiation pattern may be referred to as a beam, which may or may not be unimodal and may allow a device to amplify a signal transmitted or received from a spatial direction.
[0204] In some embodiments, the antenna panels may or may not be virtualized as antenna ports in the specification. The antenna panels may be connected to the baseband processing module via a radio frequency ("RF") chain for each of the transmit (egress) and receive (ingress) directions. The capabilities of the device in terms of the number of antenna panels, its duplexing capabilities, its beamforming capabilities, etc. may or may not be transparent to other devices. In some embodiments, the capability information may be communicated via signaling, or in some embodiments, the capability information may be provided to the device without signaling. In the case where such information is available to other devices, the information may be used for signaling or local decision making.
[0205] In some embodiments, a device (e.g., UE, node) antenna panel may be a physical or logical antenna array that includes a collection of antenna elements or antenna ports that share a common or significant portion of an RF chain (e.g., in-phase / quadrature ("I / Q") modem, analog / digital ("A / D") converter, local oscillator, phase shift network). A device antenna panel or "device panel" may be a logical entity having physical device antennas mapped to logical entities. The mapping of physical device antennas to logical entities may depend on the device implementation. Communication (reception or transmission) on at least a subset of antenna elements or antenna ports (also referred to herein as active elements) that are in an active state to radiate energy of the antenna panel requires biasing or powering the RF chain, which results in current consumption or power consumption in the device associated with the antenna panel (including power consumption of power amplifiers / low noise amplifiers ("LNA") associated with the antenna elements or antenna ports). The phrase "in an active state to radiate energy" as used herein is not meant to be limited to transmission functions, but also covers reception functions. Thus, antenna elements that are in an active state to radiate energy may be coupled simultaneously or sequentially to a transmitter to transmit RF energy or to a receiver to receive RF energy, or generally to a transceiver for performing its intended functionality. Communication on the active elements of the antenna panel enables the generation of a radiation pattern or beam.
[0206] In some embodiments, depending on the device's own implementation, the "device panel" may have at least one of the following functionalities, as an operational role of a unit for antenna groups to independently control their Tx beams, a unit for antenna groups to independently control their transmission power, a unit for antenna groups to independently control their transmission timing. The "device panel" may be transparent to the gNB. For certain conditions, the gNB or the network may assume that the mapping between the physical antennas of the device and the logical entity "device panel" cannot be changed. For example, the conditions may include until the next update or report from the device or include the duration for which the gNB assumes that the mapping will not change. The device may report its capabilities regarding the "device panel" to the gNB or the network. The device capabilities may include at least the number of "device panels". In one implementation, the device may support UL transmissions from one beam within a panel; with multiple panels, more than one beam may be used for UL transmission (one beam per panel). In another implementation, each panel may support / be used for UL transmissions more than one beam.
[0207] In some of the described embodiments, antenna ports are defined such that a channel conveying a symbol on the antenna port can be inferred from a channel conveying another symbol on the same antenna port.
[0208] Two antenna ports are said to be quasi-co-located if a large-scale property of the channel conveying symbols on one antenna port can be inferred from the channel conveying symbols on the other antenna port. The large-scale properties include one or more of delay spread, Doppler spread, Doppler shift, average gain, average delay, and spatial Rx parameters. Two antenna ports may be quasi-co-located with respect to a subset of large-scale properties, and different subsets of large-scale properties may be indicated by the QCL type. The QCL type may indicate which channel properties are the same between two reference signals (e.g., on two antenna ports). Thus, reference signals may be linked to each other with respect to what the UE may assume about its channel statistics or QCL properties. For example, qcl-Type may take one of the following values: "QCL-TypeA": {Doppler shift, Doppler spread, average delay, delay spread}, "QCL-TypeB": {Doppler shift, Doppler spread}, "QCL-TypeC": {Doppler shift, average delay} and / or "QCL-TypeD": {spatial Rx parameters}.
[0209] The spatial Rx parameters may include one or more of the following: angle of arrival (AoA), dominant AoA, average AoA, angular spread, power angular spectrum (PAS) of AoA, average AoD (angle of departure), PAS of AoD, transmit / receive channel correlation, transmit / receive beamforming, spatial channel correlation, and the like.
[0210] QCL-TypeA, QCL-TypeB, and QCL-TypeC are applicable to all carrier frequencies, but QCL-TypeD is applicable only in higher carrier frequencies (e.g., mmWave, FR2, and higher), where the UE is not inherently capable of performing omni-directional transmissions, i.e., the UE will need to form a beam for directional transmissions. For QCL-TypeD between two reference signals A and B, reference signal A is considered to be spatially co-located with reference signal B, and the UE may assume that reference signals A and B may be received using the same spatial filter (e.g., with the same RX beamforming weights).
[0211] An "antenna port" according to an embodiment may be a logical port that may correspond to a beam (caused by beamforming) or may correspond to a physical antenna on a device. In some embodiments, a physical antenna may be directly mapped to a single antenna port, where the antenna port corresponds to an actual physical antenna. Alternately, a physical antenna set or subset or an antenna set or antenna array or antenna subarray may be mapped to one or more antenna ports after applying complex weights, cyclic delays, or both to the signal on each physical antenna. A physical antenna set may have antennas from a single module or panel or from multiple modules or panels. The weights may be fixed, such as in an antenna virtualization scheme, for example, cyclic delay diversity (CDD). The process for deriving antenna ports from physical antennas may be specific to a device implementation and transparent to other devices.
[0212] In some of the described embodiments, a TCI state (transmission configuration indication) associated with a target transmission may indicate parameters for configuring a quasi-colocation relationship between a target transmission (e.g., a target RS for a DM-RS port of a target transmission during a transmission opportunity) and a source reference signal (e.g., SSB / CSI-RS / SRS) relative to a quasi-colocation type parameter indicated in the corresponding TCI state. The TCI describes which reference signals are used as QCL sources and which QCL properties can be derived from each reference signal. A device may receive a configuration of multiple transmission configuration indicator states for a serving cell for transmission on the serving cell. In some of the described embodiments, the TCI state includes at least one source RS to provide a reference (UE assumption) for determining the QCL and / or spatial filter.
[0213] In some of the described embodiments, the spatial relationship information associated with the target transmission may indicate parameters for configuring a spatial setting between the target transmission and a reference RS (e.g., SSB / CSI-RS / SRS). For example, the device may transmit the target transmission using the same spatial domain filter used to receive the reference RS (e.g., DL RS such as SSB / CSI-RS). In another example, the device may transmit the target transmission using the same spatial domain transmission filter used to transmit the reference RS (e.g., UL RS such as SRS). The device may receive one of a plurality of spatial relationship information configurations for a serving cell for transmission on the serving cell.
[0214] In some of the described embodiments, if the device is configured with separate DL / UL TCI via RRC signaling, a UL TCI state is provided. The UL TCI state may include a source reference signal that provides a reference for determining a UL spatial domain transmission filter for UL transmissions (e.g., PUSCH based on dynamic grant / configured grant, dedicated PUCCH resources) in a CC or across a set of configured CCs / BWPs.
[0215] In some of the described embodiments, if the device is configured with a joint DL / UL TCI via RRC signaling (e.g., the configuration of the joint TCI or separate DL / UL TCI is based on RRC signaling), a joint DL / UL TCI state is provided. The joint DL / UL TCI state refers to at least one common source reference RS used to determine both DL QCL information and UL spatial transmission filters. A QCL type D indication (e.g., for device-specific PDCCH / PDSCH) is provided from a source RS determined from the indicated joint (or common) TCI state and is used to determine a UL spatial transmission filter (e.g., for UE-specific PUSCH / PUCCH) for a CC or a set of cross-configured CCs / BWPs. In one example, the UL spatial transmission filter is derived from a DL QCL type D RS in the joint TCI state. The spatial setting of the UL transmission can be based on a spatial relationship with reference to a source RS configured with a qcl-Type set set to "type D" in the joint TCI state.
[0216] Figure 7 An example of an NR protocol stack 700 that supports techniques for joint CSI training and PMI feedback for AI-enabled networks is illustrated in accordance with aspects of the present disclosure. Figure 7UE, RAN node and 5G core network are shown, but these are representative of a collection of remote units 105 interacting with base unit 121 and mobile core network 140. As depicted, protocol stack 700 includes user plane protocol stack 701 and control plane protocol stack 703. User plane protocol stack 701 includes physical ("PHY") layer 705, medium access control ("MAC") sublayer 707, radio link control ("RLC") sublayer 709, packet data convergence protocol ("PDCP") sublayer 711, and service data adaptation protocol ("SDAP") layer 713. Control plane protocol stack 703 includes physical layer 705, MAC sublayer 707, RLC sublayer 709, and PDCP sublayer 711. Control plane protocol stack 703 also includes RRC layer 715 and non-access stratum ("NAS") layer 717.
[0217] The AS layer for the user plane protocol stack 701 (also referred to as the "AS protocol stack") consists of at least the SDAP sublayer, the PDCP sublayer, the RLC sublayer, the MAC sublayer, and the physical layer. The AS layer for the control plane protocol stack 703 consists of at least the RRC sublayer, the PDCP sublayer, the RLC sublayer, the MAC sublayer, and the physical layer. Layer 2 ("L2") is split into the SDAP sublayer, the PDCP sublayer, the RLC sublayer, and the MAC sublayer. Layer 3 ("L3") includes the RRC sublayer 715 and the NAS layer 717 for the control plane, and includes, for example, an Internet Protocol ("IP") layer or a PDU layer (not depicted) for the user plane. L1 and L2 are referred to as "lower layers", while L3 and above (e.g., transport layer, application layer) are referred to as "higher layers" or "upper layers".
[0218] The physical layer 705 provides transport channels to the MAC sublayer 707. The MAC sublayer 707 provides logical channels to the RLC sublayer 709. The RLC sublayer 709 provides RLC channels to the PDCP sublayer 711. The PDCP sublayer 711 provides radio bearers to the SDAP sublayer 713 and / or the RRC layer 715. The SDAP sublayer 713 provides QoS flows to the core network (e.g., 5GC). The RRC layer 715 provides carrier aggregation and / or dual connectivity addition, modification, and release. The RRC layer 715 also manages the creation, configuration, maintenance, and release of signaling radio bearers ("SRBs") and data radio bearers ("DRBs").
[0219] Figure 8An example of a UE device 800 that supports joint CSI training and PMI feedback for an AI-enabled network according to aspects of the present disclosure is illustrated. In various embodiments, the UE device 800 is used to implement one or more of the solutions described above. The UE device 800 may be an embodiment of a UE, such as a remote unit 105 and / or a UE 805 as described above. In addition, the UE device 800 may include a processor 805, a memory 810, an input device 815, an output device 820, and a transceiver 825. In some embodiments, the input device 815 and the output device 820 are combined into a single device, such as a touch screen. In a specific embodiment, the UE device 800 may not include any input device 815 and / or output device 820. In various embodiments, the UE device 800 may include one or more of the processor 805, the memory 810, and the transceiver 825, and may not include the input device 815 and / or the output device 820.
[0220] As depicted, the transceiver 825 includes at least one transmitter 830 and at least one receiver 835. Here, the transceiver 825 communicates with one or more base units 121. In addition, the transceiver 825 can support at least one network interface 840 and / or an application interface 845. The application interface 845 can support one or more APIs. The network interface 840 can support 3GPP reference points such as Uu and PC5. As understood by those skilled in the art, other network interfaces 840 can be supported.
[0221] In one embodiment, the processor 805 may include any known controller capable of executing computer-readable instructions and / or capable of performing logical operations. For example, the processor 805 may be a microcontroller, a microprocessor, a central processing unit (“CPU”), a graphics processing unit (“GPU”), an auxiliary processing unit, a field programmable gate array (“FPGA”), or a similar programmable controller. In some embodiments, the processor 805 executes instructions stored in the memory 810 to perform the methods and routines described herein. The processor 805 is communicatively coupled to the memory 810, the input device 815, the output device 820, and the transceiver 825.
[0222] In various embodiments, the processor 805 controls the UE device 800 to implement the above-described UE behaviors for the techniques of joint CSI training and PMI feedback for AI-enabled networks.
[0223] In one embodiment, memory 810 is a computer-readable storage medium. In some embodiments, memory 810 includes volatile computer storage media. For example, memory 810 may include RAM, including dynamic RAM ("DRAM"), synchronous dynamic RAM ("SDRAM"), and / or static RAM ("SRAM"). In some embodiments, memory 810 includes non-volatile computer storage media. For example, memory 810 may include a hard drive, flash memory, or any other suitable non-volatile computer storage device. In some embodiments, memory 810 includes both volatile computer storage media and non-volatile computer storage media.
[0224] In some embodiments, memory 810 stores data related to techniques for joint CSI training and PMI feedback for AI-enabled networks. For example, memory 810 may store UL, DL, and / or SL resource configurations, measurement configurations, UE configurations, beam management states, etc. In a particular embodiment, memory 810 also stores programming code and related data, such as an operating system or other controller algorithm operating on remote unit 135.
[0225] In one embodiment, the input device 815 may include any known computer input device, including a touch panel, buttons, keyboard, stylus, microphone, etc. In some embodiments, the input device 815 may be integrated with the output device 820, such as a touch screen or similar touch-sensitive display. In some embodiments, the input device 815 includes a touch screen so that text can be entered using a virtual keyboard displayed on the touch screen and / or by handwriting on the touch screen. In some embodiments, the input device 815 includes two or more different devices, such as a keyboard and a touch panel.
[0226] In one embodiment, the output device 820 is designed to output visual, auditory and / or tactile signals. In some embodiments, the output device 820 includes an electronically controllable display or display device capable of outputting visual data to a user. For example, the output device 820 may include, but is not limited to, an LCD display, an LED display, an OLED display, a projector, or a similar display device capable of outputting images, text, etc. to a user. As another non-limiting example, the output device 820 may include a wearable display that is separate from the rest of the UE device 800 but communicatively coupled to the rest of the UE device 800, such as a smart watch, smart glasses, a head-up display, and the like. In addition, the output device 820 may be a component of a smart phone, a personal digital assistant, a television, a tablet computer, a notebook (laptop) computer, a personal computer, a vehicle dashboard, and the like.
[0227] In certain embodiments, the output device 820 includes one or more speakers for producing sounds. For example, the output device 820 may produce an audible alarm or notification (e.g., a beep or ring tone). In some embodiments, the output device 820 includes one or more haptic devices for producing vibration, motion, or other tactile feedback. In some embodiments, all or part of the output device 820 may be integrated with the input device 815. For example, the input device 815 and the output device 820 may form a touch screen or similar touch-sensitive display. In other embodiments, the output device 820 may be located near the input device 815.
[0228] The transceiver 825 includes at least a transmitter 830 and at least one receiver 835. The transceiver 825 may be used to provide UL communication signals to the base unit 121 and receive DL communication signals from the base unit 121, as described herein. Similarly, the transceiver 825 may be used to transmit and receive SL signals (e.g., V2X communication), as described herein. Although only one transmitter 830 and one receiver 835 are illustrated, the UE device 800 may have any suitable number of transmitters 830 and receivers 835. Furthermore, the transmitter 830 and the receiver 835 may be any suitable type of transmitter and receiver. In one embodiment, the transceiver 825 includes a first transmitter / receiver pair for communicating with a mobile communication network on a licensed radio spectrum and a second transmitter / receiver pair for communicating with a mobile communication network on an unlicensed radio spectrum.
[0229] In a particular embodiment, a first transmitter / receiver pair for communicating with a mobile communication network on a licensed radio spectrum and a second transmitter / receiver pair for communicating with a mobile communication network on an unlicensed radio spectrum may be combined into a single transceiver unit, such as a single chip, that performs functions for use with both the licensed radio spectrum and the unlicensed radio spectrum. In some embodiments, the first transmitter / receiver pair and the second transmitter / receiver pair may share one or more hardware components. For example, a particular transceiver 825, transmitter 830, and receiver 835 may be implemented as physically separate components that access shared hardware resources and / or software resources, such as a network interface 840.
[0230] In various embodiments, one or more transmitters 830 and / or one or more receivers 835 may be implemented and / or integrated as a single hardware component, such as a multi-transceiver chip, a system on a chip, an ASIC, or other type of hardware component. In a particular embodiment, one or more transmitters 830 and / or one or more receivers 835 may be implemented and / or integrated as a multi-chip module. In some embodiments, other components (e.g., a network interface 840) or other hardware components / circuits may be integrated into a single chip with any number of transmitters 830 and / or receivers 835. In such embodiments, the transmitters 830 and receivers 835 may be logically configured as transceivers 825 using one or more common control signals or as modular transmitters 830 and receivers 835 implemented in the same hardware chip or in a multi-chip module.
[0231] In one embodiment, the memory 810 includes instructions that can be executed by the processor 805 to cause the device 800 to receive a set of RSs for channel measurement corresponding to NZP CSI-RS resources based on the CSI report setting, generate a CSI report including precoding matrix information for a precoding matrix and a set of coefficients corresponding to at least one dimension of the precoding matrix based on the channel measurement, and transmit the CSI report.
[0232] In one embodiment, the precoding matrix includes a first basis transform associated with a first dimension of at least one dimension of the precoding matrix, and the set of coefficients corresponds to the precoding matrix decomposed into two coefficient subsets, the two coefficient subsets corresponding to different indices of the first dimension of the precoding matrix.
[0233] In one embodiment, a first subset of the two coefficient subsets associated with the second basis transform is associated with a second dimension of the precoding matrix, and a second subset of the two coefficient subsets associated with the third basis transform is associated with the second dimension of the precoding matrix.
[0234] In one embodiment, the instructions are executable by the processor 805 to cause the apparatus 800 to receive, from the network, a CSI reporting setting for configuring the UE for at least one of channel measurement and reporting.
[0235] In one embodiment, the NZP CSI-RS resource includes a plurality of CSI-RS ports.
[0236] In one embodiment, the third basis transform transforms the second dimension of the precoding matrix from a domain including multiple CSI-RS ports to a domain including one or more spatial beams.
[0237] In one embodiment, the instructions are executable by the processor 805 to cause the apparatus 800 to transform a first dimension of the precoding matrix from a domain including multiple CSI-RS ports to a domain including one or more spatial beams based on a first basis transform.
[0238] In one embodiment, the second basis transform includes at least one of: averaging multiple indices of the second dimension into a fewer number of indices corresponding to the second dimension, a one-to-one transform into the same number of indices corresponding to the second dimension, and selecting a subset of indices from a set of indices corresponding to the second dimension.
[0239] In one embodiment, the second basis transform includes selecting one of an even index or an odd index corresponding to the second dimension.
[0240] In one embodiment, the second basis transform corresponds to a third basis transform of the same type, the third basis transform transforming the second dimension into a smaller number of indices than the second basis transform.
[0241] In one embodiment, the third basis transform corresponds to at least one of: transforming the second dimension of the precoding matrix from a domain comprising frequency subbands of multiple channels to a domain comprising a transformed frequency domain basis, and transforming the second dimension of the precoding matrix from a domain comprising time units of multiple channels to a domain comprising a transformed time domain basis.
[0242] In one embodiment, the first basis transform corresponds to transforming the first dimension of the precoding matrix according to at least one of the following transformation modes: transforming the first dimension of the precoding matrix from a domain including frequency subbands of multiple channels to a domain including a transformed frequency domain basis and transforming the first dimension of the precoding matrix from a domain including time units of multiple channels to a domain including a transformed time domain basis.
[0243] In one embodiment, the instructions are executable by the processor 805 to cause the apparatus 800 to represent coefficients associated with the first subset of coefficients with a different codebook of quantized values than coefficients associated with the second subset of coefficients.
[0244] In one embodiment, instructions are executable by processor 805 to cause apparatus 800 to separately quantize coefficients associated with the first subset of coefficients for each of an amplitude component and a phase component of the coefficients.
[0245] In one embodiment, instructions are executable by processor 805 to cause apparatus 800 to jointly quantize a coefficient associated with the second subset of coefficients for both an amplitude component and a phase component of the coefficient.
[0246] In one embodiment, the instructions are executable by the processor 805 to cause the apparatus 800 to jointly quantize a plurality of coefficients associated with the second subset of coefficients for at least one of an amplitude component and a phase component.
[0247] In one embodiment, the number of bits associated with quantization and coefficients associated with the first subset of coefficients is greater than the number of bits associated with quantization and coefficients associated with the second subset of coefficients.
[0248] In one embodiment, instructions are executable by processor 805 to cause apparatus 800 to represent coefficients associated with the second subset of coefficients using a codebook from the set of preconfigured codebooks.
[0249] In one embodiment, the set of preconfigured codebooks includes at least one of a codebook having amplitude values uniformly distributed in a linear domain and a codebook having amplitude values uniformly distributed in a logarithmic domain.
[0250] In one embodiment, the instructions are executable by the processor 805 to cause the apparatus 800 to configure a third basis transform associated with the second subset of coefficients according to a set of preconfigured transform methods.
[0251] In one embodiment, the set of preconfigured transform methods includes at least one of a discrete Fourier transform, a discrete cosine transform, a discrete sine transform, a discrete wavelet transform, and a Haar transform.
[0252] In one embodiment, the instructions may be executed by the processor 805 to cause the apparatus 800 to feedback an index corresponding to a vector of a transform matrix corresponding to a third transform in the CSI report.
[0253] In one embodiment, the network configures, via a higher layer configuration, an index corresponding to a vector of a transformation matrix corresponding to the third transformation.
[0254] In one embodiment, the instructions are executable by the processor 805 to cause the apparatus 800 to configure the UE to report a reporting amount corresponding to the training indicator.
[0255] In one embodiment, the CSI report includes location information corresponding to the location of the UE.
[0256] In one embodiment, the CSI report includes a bitmap indicating indices of non-zero coefficients, and wherein coefficients corresponding to the first subset of channel coefficients are excluded from the bitmap.
[0257] In one embodiment, coefficients corresponding to the first of the two subsets of channel coefficients are always reported in the CSI report.
[0258] In one embodiment, the instructions are executable by the processor 805 to cause the apparatus 800 to configure the UE with a higher layer parameter within a CSI reporting setting that enables or disables reporting of a first subset of the two coefficient subsets.
[0259] Fig. 9 An example of a NE device 900 that supports techniques for joint CSI training and PMI feedback for AI-enabled networks according to aspects of the present disclosure is illustrated. In some embodiments, the NE device 900 may be an embodiment of a RAN node and its supporting hardware, such as the base unit 121 and / or gNB described above. In addition, the NE device 900 may include a processor 905, a memory 910, an input device 915, an output device 920, and a transceiver 925. In a specific embodiment, the NE device 900 does not include any input device 915 and / or output device 920.
[0260] As depicted, the transceiver 925 includes at least one transmitter 930 and at least one receiver 935. Here, the transceiver 925 communicates with one or more remote units 105. In addition, the transceiver 925 can support at least one network interface 940 and / or application interface 945. The application interface 945 can support one or more APIs. The network interface 940 can support 3GPP reference points, such as Uu, N1, N2, and N3 interfaces. As understood by those skilled in the art, other network interfaces 940 can be supported.
[0261] In one embodiment, the processor 905 may include any known controller capable of executing computer-readable instructions and / or capable of performing logical operations. For example, the processor 905 may be a microcontroller, a microprocessor, a CPU, a GPU, an auxiliary processing unit, an FPGA, or a similar programmable controller. In some embodiments, the processor 905 executes instructions stored in the memory 910 to perform the methods and routines described herein. The processor 905 is communicatively coupled to the memory 910, the input device 915, the output device 920, and the transceiver 925.
[0262] In various embodiments, the processor 905 controls the NE device 900 to implement the above-described network entity behavior (e.g., of the gNB) for the technique of joint CSI training and PMI feedback for an AI-enabled network.
[0263] In one embodiment, the memory 910 is a computer-readable storage medium. In some embodiments, the memory 910 includes volatile computer storage media. For example, the memory 910 may include RAM, which includes DRAM, SDRAM and / or SRAM. In some embodiments, the memory 910 includes non-volatile computer storage media. For example, the memory 910 may include a hard drive, flash memory, or any other suitable non-volatile computer storage device. In some embodiments, the memory 910 includes both volatile computer storage media and non-volatile computer storage media.
[0264] In some embodiments, the memory 910 stores data related to techniques for joint CSI training and PMI feedback for AI-enabled networks. For example, the memory 910 may store UL, DL, and / or SL resource configurations, measurement configurations, UE configurations, beam management states, etc. In a particular embodiment, the memory 910 also stores programming code and related data, such as an operating system ("OS") or other controller algorithms operating on the NE device 900 and one or more software applications.
[0265] In one embodiment, the input device 915 may include any known computer input device, including a touch panel, buttons, a keyboard, a stylus, a microphone, and the like. In some embodiments, the input device 915 may be integrated with the output device 920, such as a touch screen or similar touch-sensitive display. In some embodiments, the input device 915 includes a touch screen so that text can be entered using a virtual keyboard displayed on the touch screen and / or by handwriting on the touch screen. In some embodiments, the input device 915 includes two or more different devices, such as a keyboard and a touch panel.
[0266] In one embodiment, the output device 920 may include any known electronically controllable display or display device. The output device 920 may be designed to output visual, auditory, and / or tactile signals. In some embodiments, the output device 920 includes an electronic display capable of outputting visual data to a user. In addition, the output device 920 may be a component of a smart phone, a personal digital assistant, a television, a tablet computer, a notebook (laptop) computer, a personal computer, a vehicle dashboard, etc.
[0267] In certain embodiments, the output device 920 includes one or more speakers for producing sounds. For example, the output device 920 may produce an audible alarm or notification (e.g., a beep or ring tone). In some embodiments, the output device 920 includes one or more haptic devices for producing vibration, motion, or other tactile feedback. In some embodiments, all or part of the output device 920 may be integrated with the input device 915. For example, the input device 915 and the output device 920 may form a touch screen or similar touch-sensitive display. In other embodiments, all or part of the output device 920 may be located near the input device 915.
[0268] As discussed above, the transceiver 925 may communicate with one or more remote units and / or with one or more networking functions that provide access to one or more PLMNs. The transceiver 925 may also communicate with one or more network functions (e.g., in the mobile core network 140). The transceiver 925 operates under the control of the processor 905 to transmit messages, data, and other signals and also receive messages, data, and other signals. For example, the processor 905 may selectively activate the transceiver (or portions thereof) at specific times in order to send and receive messages.
[0269] The transceiver 925 may include one or more transmitters 930 and one or more receivers 935. In a particular embodiment, the one or more transmitters 930 and / or the one or more receivers 935 may share transceiver hardware and / or circuitry. For example, the one or more transmitters 930 and / or the one or more receivers 935 may share antennas, antenna tuners, amplifiers, filters, oscillators, mixers, modulators / demodulators, power supplies, etc. In one embodiment, the transceiver 925 implements multiple logical transceivers using different communication protocols or protocol stacks while using common physical hardware.
[0270] In one embodiment, the memory 910 includes instructions that can be executed by the processor 905 to transmit to the UE a CSI report setting for configuring the UE for at least one of channel measurement and reporting, transmit to the UE a set of RSs for channel measurement corresponding to NZP CSI-RS resources, and receive a CSI report from the UE. In one embodiment, the CSI report includes precoding matrix information for a precoding matrix and a set of coefficients corresponding to at least one dimension of the precoding matrix, the precoding matrix including a first basis transform associated with a first dimension of the at least one dimension of the precoding matrix, the set of coefficients corresponding to the precoding matrix decomposed into two coefficient subsets, the two coefficient subsets corresponding to different indices of the first dimension of the precoding matrix, the first subset of the two coefficient subsets being associated with a second basis transform, the second basis transform being associated with a second dimension of the precoding matrix, and the second subset of the two coefficient subsets being associated with a third basis transform, the third basis transform being associated with the second dimension of the precoding matrix.
[0271] Fig.10 A flow chart illustrating a method 1000 of techniques for supporting joint CSI training and PMI feedback for an AI-enabled network in accordance with aspects of the present disclosure. The method 1000 may be performed by a UE (e.g., remote unit 105 and / or UE device 800) as described herein. In some embodiments, the method 1000 may be performed by a processor (e.g., a microcontroller, microprocessor, CPU, GPU, auxiliary processing unit, FPGA, etc.) executing programming code.
[0272] In one embodiment, method 1000 starts and receives 1005 a set of RSs for channel measurement corresponding to NZP CSI-RS resources based on CSI report settings. In one embodiment, method 1000 generates 1010 a CSI report based on the channel measurement, the CSI report including precoding matrix information for a precoding matrix and a set of coefficients corresponding to at least one dimension of the precoding matrix. In one embodiment, method 1000 transmits 1015 the CSI report, and method 1000 ends.
[0273] Fig.11 A flow chart illustrating a method 1100 of techniques for supporting joint CSI training and PMI feedback for AI-enabled networks according to aspects of the present disclosure. The method 1100 may be performed by a network device (e.g., base unit 121 and / or NE device 900) as described herein. In some embodiments, the method 1100 may be performed by a processor (e.g., a microcontroller, a microprocessor, a CPU, a GPU, an auxiliary processing unit, an FPGA, etc.) executing programming code.
[0274] In one embodiment, the method 1100 starts and transmits 1105 a CSI report setting to a UE, the CSI report setting being used to configure the UE for at least one of channel measurement and reporting. In one embodiment, the method 1100 transmits 1110 to the UE a set of RSs for channel measurement corresponding to NZP CSI-RS resources. In one embodiment, the method 1100 receives 1115 a CSI report from the UE, the CSI report including precoding matrix information for a precoding matrix and a set of coefficients corresponding to at least one dimension of the precoding matrix, and the method 1100 ends.
[0275] A first device for joint CSI training and PMI feedback for AI-enabled networks is disclosed. The first device may include a UE as described herein, such as a remote unit 105 and / or a UE device 800. In some embodiments, the first device may include a processor that executes programming code, such as a microcontroller, a microprocessor, a CPU, a GPU, an auxiliary processing unit, an FPGA, etc.
[0276] In one embodiment, the first apparatus includes a processor and a memory coupled to the processor. The memory includes instructions executable by the processor to cause the apparatus to receive a set of RSs for channel measurement corresponding to NZP CSI-RS resources based on a CSI report setting, generate a CSI report based on the channel measurement, and transmit the CSI report, wherein the CSI report includes precoding matrix information for a precoding matrix and a set of coefficients corresponding to at least one dimension of the precoding matrix.
[0277] In one embodiment, the precoding matrix includes a first basis transform associated with a first dimension of at least one dimension of the precoding matrix, and the set of coefficients corresponds to the precoding matrix decomposed into two coefficient subsets, the two coefficient subsets corresponding to different indices of the first dimension of the precoding matrix.
[0278] In one embodiment, a first subset of the two coefficient subsets associated with the second basis transform is associated with a second dimension of the precoding matrix, and a second subset of the two coefficient subsets associated with the third basis transform is associated with the second dimension of the precoding matrix.
[0279] In one embodiment, the instructions are executable by the processor to cause the apparatus to receive a CSI reporting setting from a network, the CSI reporting setting being used to configure the UE for at least one of channel measurement and reporting.
[0280] In one embodiment, the NZP CSI-RS resource includes a plurality of CSI-RS ports.
[0281] In one embodiment, the third basis transform transforms the second dimension of the precoding matrix from a domain including multiple CSI-RS ports to a domain including one or more spatial beams.
[0282] In one embodiment, the instructions are executable by the processor to cause the apparatus to transform a first dimension of the precoding matrix from a domain including multiple CSI-RS ports to a domain including one or more spatial beams based on a first basis transform.
[0283] In one embodiment, the second basis transform includes at least one of: averaging multiple indices of the second dimension into a fewer number of indices corresponding to the second dimension, a one-to-one transform into the same number of indices corresponding to the second dimension, and selecting a subset of indices from a set of indices corresponding to the second dimension.
[0284] In one embodiment, the second basis transform includes selecting one of an even index or an odd index corresponding to the second dimension.
[0285] In one embodiment, the second basis transform corresponds to a third basis transform of the same type, the third basis transform transforming the second dimension into a smaller number of indices than the second basis transform.
[0286] In one embodiment, the third basis transform corresponds to at least one of: transforming the second dimension of the precoding matrix from a domain comprising frequency subbands of multiple channels to a domain comprising a transformed frequency domain basis, and transforming the second dimension of the precoding matrix from a domain comprising time units of multiple channels to a domain comprising a transformed time domain basis.
[0287] In one embodiment, the first basis transform corresponds to transforming the first dimension of the precoding matrix according to at least one of the following transformation modes: transforming the first dimension of the precoding matrix from a domain including frequency subbands of multiple channels to a domain including a transformed frequency domain basis and transforming the first dimension of the precoding matrix from a domain including time units of multiple channels to a domain including a transformed time domain basis.
[0288] In one embodiment, the instructions are executable by the processor to cause the device to represent coefficients associated with the first subset of coefficients with a different codebook of quantized values than coefficients associated with the second subset of coefficients.
[0289] In one embodiment, the instructions are executable by the processor to cause the apparatus to separately quantize coefficients associated with the first subset of coefficients for each of an amplitude component and a phase component of the coefficients.
[0290] In one embodiment, the instructions are executable by the processor to cause the apparatus to jointly quantize a coefficient associated with the second subset of coefficients for both an amplitude component and a phase component of the coefficient.
[0291] In one embodiment, the instructions are executable by the processor to cause the apparatus to jointly quantize a plurality of coefficients associated with the second subset of coefficients for at least one of an amplitude component and a phase component.
[0292] In one embodiment, the number of bits associated with quantization and coefficients associated with the first subset of coefficients is greater than the number of bits associated with quantization and coefficients associated with the second subset of coefficients.
[0293] In one embodiment, the instructions are executable by the processor to cause the apparatus to represent coefficients associated with the second subset of coefficients utilizing a codebook from the set of preconfigured codebooks.
[0294] In one embodiment, the set of preconfigured codebooks includes at least one of a codebook having amplitude values uniformly distributed in a linear domain and a codebook having amplitude values uniformly distributed in a logarithmic domain.
[0295] In one embodiment, the instructions are executable by the processor to cause the apparatus to configure a third basis transform associated with the second subset of coefficients according to a set of preconfigured transform methods.
[0296] In one embodiment, the set of preconfigured transform methods includes at least one of a discrete Fourier transform, a discrete cosine transform, a discrete sine transform, a discrete wavelet transform, and a Haar transform.
[0297] In one embodiment, the instructions are executable by the processor to cause the device to feedback an index corresponding to a vector of a transform matrix corresponding to a third transform in the CSI report.
[0298] In one embodiment, the network configures, via a higher layer configuration, an index corresponding to a vector of a transformation matrix corresponding to the third transformation.
[0299] In one embodiment, the instructions are executable by the processor to cause the apparatus to configure the UE to report a reporting amount corresponding to the training indicator.
[0300] In one embodiment, the CSI report includes location information corresponding to the location of the UE.
[0301] In one embodiment, the CSI report includes a bitmap indicating indices of non-zero coefficients, and wherein coefficients corresponding to the first subset of channel coefficients are excluded from the bitmap.
[0302] In one embodiment, coefficients corresponding to the first of the two subsets of channel coefficients are always reported in the CSI report.
[0303] In one embodiment, the instructions are executable by the processor to cause the apparatus to configure the UE with a higher layer parameter within a CSI reporting setting that enables or disables reporting of a first subset of the two coefficient subsets.
[0304] A first method of techniques for joint CSI training and PMI feedback for an AI-enabled network is disclosed. The first method may be performed by a UE (e.g., a remote unit 105 and / or a UE device 800) as described herein. In some embodiments, the first method may be performed by a processor (e.g., a microcontroller, a microprocessor, a CPU, a GPU, an auxiliary processing unit, an FPGA, etc.) executing programming code.
[0305] In one embodiment, the first method receives a set of RSs for channel measurement corresponding to NZP CSI-RS resources based on a CSI report setting, and generates a CSI report based on the channel measurement, wherein the CSI report includes precoding matrix information for a precoding matrix and a set of coefficients corresponding to at least one dimension of the precoding matrix. In one embodiment, the first method transmits the CSI report.
[0306] In one embodiment, the precoding matrix includes a first basis transform associated with a first dimension of at least one dimension of the precoding matrix, and the set of coefficients corresponds to the precoding matrix decomposed into two coefficient subsets, the two coefficient subsets corresponding to different indices of the first dimension of the precoding matrix.
[0307] In one embodiment, a first subset of the two coefficient subsets associated with the second basis transform is associated with a second dimension of the precoding matrix, and a second subset of the two coefficient subsets associated with the third basis transform is associated with the second dimension of the precoding matrix.
[0308] In one embodiment, a first method receives a CSI reporting setting from a network for configuring a UE for at least one of channel measurement and reporting.
[0309] In one embodiment, the NZP CSI-RS resource includes a plurality of CSI-RS ports.
[0310] In one embodiment, the third basis transform transforms the second dimension of the precoding matrix from a domain including multiple CSI-RS ports to a domain including one or more spatial beams.
[0311] In one embodiment, the first method transforms a first dimension of a precoding matrix from a domain including a plurality of CSI-RS ports to a domain including one or more spatial beams based on a first basis transform.
[0312] In one embodiment, the second basis transform includes at least one of: averaging multiple indices of the second dimension into a fewer number of indices corresponding to the second dimension, a one-to-one transform into the same number of indices corresponding to the second dimension, and selecting a subset of indices from a set of indices corresponding to the second dimension.
[0313] In one embodiment, the second basis transform includes selecting one of an even index or an odd index corresponding to the second dimension.
[0314] In one embodiment, the second basis transform corresponds to a third basis transform of the same type, the third basis transform transforming the second dimension into a smaller number of indices than the second basis transform.
[0315] In one embodiment, the third basis transform corresponds to at least one of: transforming the second dimension of the precoding matrix from a domain comprising frequency subbands of multiple channels to a domain comprising a transformed frequency domain basis, and transforming the second dimension of the precoding matrix from a domain comprising time units of multiple channels to a domain comprising a transformed time domain basis.
[0316] In one embodiment, the first basis transform corresponds to transforming the first dimension of the precoding matrix according to at least one of the following transformation modes: transforming the first dimension of the precoding matrix from a domain including frequency subbands of multiple channels to a domain including a transformed frequency domain basis and transforming the first dimension of the precoding matrix from a domain including time units of multiple channels to a domain including a transformed time domain basis.
[0317] In one embodiment, the first method represents coefficients associated with the first subset of coefficients with a different codebook of quantized values than coefficients associated with the second subset of coefficients.
[0318] In one embodiment, a first method quantizes coefficients associated with a first subset of coefficients separately for each of the amplitude component and the phase component of the coefficients.
[0319] In one embodiment, a first method jointly quantizes a coefficient associated with the second subset of coefficients for both an amplitude component and a phase component of the coefficient.
[0320] In one embodiment, the first method jointly quantizes a plurality of coefficients associated with the second subset of coefficients for at least one of an amplitude component and a phase component.
[0321] In one embodiment, the number of bits associated with quantization and coefficients associated with the first subset of coefficients is greater than the number of bits associated with quantization and coefficients associated with the second subset of coefficients.
[0322] In one embodiment, a first method represents coefficients associated with the second subset of coefficients using a codebook from a set of preconfigured codebooks.
[0323] In one embodiment, the set of preconfigured codebooks includes at least one of a codebook having amplitude values uniformly distributed in a linear domain and a codebook having amplitude values uniformly distributed in a logarithmic domain.
[0324] In one embodiment, the first method configures a third basis transform associated with the second subset of coefficients according to a set of preconfigured transform methods.
[0325] In one embodiment, the set of preconfigured transform methods includes at least one of a discrete Fourier transform, a discrete cosine transform, a discrete sine transform, a discrete wavelet transform, and a Haar transform.
[0326] In one embodiment, the first method feeds back an index corresponding to a vector of a transformation matrix corresponding to a third transformation in the CSI report.
[0327] In one embodiment, the network configures, via a higher layer configuration, an index corresponding to a vector of a transformation matrix corresponding to the third transformation.
[0328] In one embodiment, a first method configures the UE to report a reporting amount corresponding to a training indicator.
[0329] In one embodiment, the CSI report includes location information corresponding to the location of the UE.
[0330] In one embodiment, the CSI report includes a bitmap indicating indices of non-zero coefficients, and wherein coefficients corresponding to the first subset of channel coefficients are excluded from the bitmap.
[0331] In one embodiment, coefficients corresponding to the first of the two subsets of channel coefficients are always reported in the CSI report.
[0332] In one embodiment, a first method configures the UE with a higher layer parameter within a CSI reporting setting that enables or disables reporting of a first subset of the two coefficient subsets.
[0333] A second device for joint CSI training and PMI feedback for AI-enabled networks is disclosed. The second device may include a network device as described herein, such as a base unit 121 and / or a NE device 900. In some embodiments, the second device may include a processor that executes programming code, such as a microcontroller, a microprocessor, a CPU, a GPU, an auxiliary processing unit, an FPGA, etc.
[0334] In one embodiment, the second device includes a processor and a memory coupled to the processor. The memory includes instructions that can be executed by the processor to cause the device to transmit to the UE a CSI report setting for configuring the UE for at least one of channel measurement and reporting, transmit to the UE an RS set for channel measurement corresponding to an NZP CSI-RS resource, and receive a CSI report from the UE. In one embodiment, the CSI report includes precoding matrix information for a precoding matrix and a coefficient set corresponding to at least one dimension of the precoding matrix, the precoding matrix includes a first basis transform associated with a first dimension of at least one dimension of the precoding matrix, the coefficient set corresponds to a precoding matrix decomposed into two coefficient subsets, the two coefficient subsets correspond to different indices of the first dimension of the precoding matrix, the first subset of the two coefficient subsets is associated with a second basis transform, the second basis transform is associated with a second dimension of the precoding matrix, and the second subset of the two coefficient subsets is associated with a third basis transform, the third basis transform is associated with the second dimension of the precoding matrix.
[0335] A second method of the technology for joint CSI training and PMI feedback for AI-enabled networks is disclosed. The second method may be performed by a network device (e.g., base unit 121 and / or NE device 900) as described herein. In some embodiments, the second method may be performed by a processor (e.g., microcontroller, microprocessor, CPU, GPU, auxiliary processing unit, FPGA, etc.) executing programming code.
[0336] In one embodiment, the second method transmits a CSI report setting to the UE, the CSI report setting being used to configure the UE for at least one of channel measurement and reporting. In one embodiment, the second method transmits a set of RSs for channel measurement corresponding to an NZP CSI-RS resource to the UE. In one embodiment, the second method receives a CSI report from the UE, the CSI report comprising precoding matrix information for a precoding matrix and a set of coefficients corresponding to at least one dimension of the precoding matrix, the precoding matrix comprising a first basis transform associated with a first dimension of at least one dimension of the precoding matrix, the set of coefficients corresponding to a precoding matrix decomposed into two coefficient subsets, the two coefficient subsets corresponding to different indices of the first dimension of the precoding matrix, the first subset of the two coefficient subsets being associated with a second basis transform, the second basis transform being associated with a second dimension of the precoding matrix, and the second subset of the two coefficient subsets being associated with a third basis transform, the third basis transform being associated with the second dimension of the precoding matrix.
[0337] Embodiments may be practiced in other specific forms. The described embodiments should be considered in all respects as merely illustrative and not restrictive. Therefore, the scope of the present invention is indicated by the appended claims rather than by the above description. All changes within the meaning and scope of the equivalent content of the claims will be included in the scope of the claims.
[0338] Those skilled in the art will appreciate that aspects of the embodiments may be embodied as systems, devices, methods, or programming products. Thus, the embodiments may be in the form of a fully hardware embodiment, a fully software embodiment (including firmware, resident software, microcode, etc.), or an embodiment combining software and hardware aspects.
[0339] For example, the disclosed embodiments may be implemented as hardware circuits including custom very large scale integration ("VLSI") circuits or gate arrays, off-the-shelf semiconductors such as logic chips, transistors or other discrete components. The disclosed embodiments may also be implemented in programmable hardware devices such as field programmable gate arrays, programmable array logic, programmable logic devices, etc. As another example, the disclosed embodiments may include one or more physical or logical blocks of executable code, which may be organized, for example, as an object, procedure, or function.
[0340] In addition, embodiments may be in the form of a programming product embodied in one or more computer-readable storage devices storing machine-readable code, computer-readable code, and / or programming code (hereinafter referred to as code). The storage device may be tangible, non-transitory, and / or non-transmitting. The storage device may not embody signals. In a particular embodiment, the storage device employs only signals for accessing the code.
[0341] Any combination of one or more computer readable media may be utilized. The computer readable medium may be a computer readable storage medium. The computer readable storage medium may be a storage device storing code. The storage device may be, for example but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, holographic, micromechanical or semiconductor system, apparatus or device, or any suitable combination of the foregoing.
[0342] More specific examples of storage devices (a non-exhaustive list) would include the following: an electrical connection having one or more conductors, a portable computer diskette, a hard disk, a random access memory ("RAM"), a read-only memory ("ROM"), an erasable programmable read-only memory ("EPROM" or flash memory), a portable compact disc read-only memory ("CD-ROM"), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a "computer-readable medium" may be any tangible medium that can contain or store programming for use by or in connection with an instruction execution system, apparatus, or device.
[0343] The code for performing the operations of the embodiments may be any number of lines written in any combination of one or more programming languages, including object-oriented programming languages (e.g., Python, Ruby, Java, Smalltalk, C++, etc.) and conventional procedural programming languages (e.g., "C" programming language, etc.) and / or machine languages (e.g., assembly language). The code may be executed entirely on the user's computer, partially on the user's computer, partially on the user's computer as a stand-alone software package, and partially on a remote computer, or entirely on a remote computer or server. In the latter scenario, the remote computer may be connected to the user's computer through any type of network including a local area network ("LAN") or a wide area network ("WAN") or may be connected to an external computer (e.g., through the Internet using an Internet service provider).
[0344] In addition, the described features, structures or characteristics of the embodiments can be combined in any suitable manner. In the following description, numerous specific details (e.g., examples of programming, software modules, user selections, network transactions, database queries, database structures, hardware modules, hardware circuits, hardware chips, etc.) are provided to provide a thorough understanding of the embodiments. However, those skilled in the art will recognize that the embodiments can be practiced without one or more of the specific details or using other methods, components, materials, etc. In other examples, well-known structures, materials, or operations are not shown or described in detail to avoid blurring aspects of the embodiments.
[0345] Throughout this specification, references to "one embodiment," "an embodiment," or similar language mean that a particular feature, structure, or characteristic described in conjunction with the embodiment is included in at least one embodiment. Therefore, throughout this specification, the phrases "in one embodiment," "in an embodiment," and similar language may, but do not necessarily, all refer to the same embodiment, but mean "one or more but not all embodiments," unless otherwise expressly specified. The terms "including," "comprising," "having," and variations thereof mean "including but not limited to," unless otherwise expressly specified. An enumerated list of items does not imply that any or all of the items are mutually exclusive, unless otherwise expressly specified. The terms "a," "an," and "the" also refer to "one or more," unless otherwise expressly specified.
[0346] As used herein, a list utilizing the conjunction "and / or" includes any single item in the list or a collection of items in the list. For example, a list containing A, B, and / or C includes only A, only B, only C, a combination of A and B, a combination of B and C, a combination of A and C, or a combination of A, B, and C. As used herein, a list using the term "one or more of" includes any single item in the list or a collection of items in the list. For example, one or more of A, B, and C includes only A, only B, only C, a combination of A and B, a combination of B and C, a combination of A and C, or a combination of A, B, and C. As used herein, a list using the term "one of" includes one and only one of any single item in the list. For example, "one of A, B, and C" includes only A, only B, or only C and does not include a combination of A, B, and C. As used herein, "a member selected from the group consisting of A, B, and C" includes one and only one of A, B, or C, and does not include a combination of A, B, and C. As used herein, "a member selected from the group consisting of A, B, and C, and a combination thereof" includes only A, only B, only C, a combination of A and B, a combination of B and C, a combination of A and C, or a combination of A, B, and C.
[0347] Aspects of the embodiments are described below with reference to schematic flow charts and / or schematic block diagrams of methods, devices, systems, and programming products according to the embodiments. It will be understood that each block of the schematic flow charts and / or schematic block diagrams and the combination of several blocks in the schematic flow charts and / or schematic block diagrams can be implemented by code. This code can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device to produce a machine, so that the instructions executed via the processor of the computer or other programmable data processing device create means for implementing the functions / actions specified in the flow charts and / or block diagrams.
[0348] This code may also be stored in a storage device that can direct a computer, other programmable data processing apparatus, or other device to function in a specific manner so that the instructions stored in the storage device produce an article of manufacture that includes instructions for implementing the functions / actions specified in the flowchart and / or block diagram.
[0349] This code may also be loaded onto a computer, other programmable data processing device, or other apparatus to initiate a series of operational steps to be performed on the computer, other programmable device, or other apparatus to produce a computer-implemented process, such that the code executed on the computer or other programmable device provides a process for implementing the functions / actions specified in the flowchart and / or block diagram.
[0350] Therefore, the flowcharts and / or block diagrams in the figures illustrate the architecture, functionality and operation of possible implementations of devices, systems, methods and programming products according to various embodiments. In this regard, each block in the flowchart or block diagram may represent a module, segment or code portion, which includes one or more executable instructions of the code for implementing the specified logical function.
[0351] It should also be noted that in some alternative implementations, the functions noted in the blocks may not occur in the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially simultaneously, or the blocks may sometimes be executed in the reverse order, depending on the functionality involved. Other steps and methods are conceivable that are equivalent in function, logic, or effect to one or more blocks, or portions thereof, of the illustrated figures.
[0352] Although various arrow types and line types may be employed in flow charts and / or block diagrams, the arrow types and line types will be understood as not limiting the scope of the corresponding embodiments. In fact, some arrows or other lines may be used to indicate only the logical flow of the depicted embodiments. For example, arrows may indicate waiting periods or monitoring periods of unspecified duration between enumerated steps of the depicted embodiments. It should also be noted that each block in the block diagrams and / or flow charts and the combination of several blocks in the block diagrams and / or flow charts may be implemented by a dedicated hardware-based system or a combination of dedicated hardware and code that performs a specified function or action.
[0353] The description of the elements in each figure can refer to the elements of the previous figure. In all figures, similar symbols refer to similar elements, including alternative embodiments of similar elements.
Claims
1. A device comprising: processor; and a memory coupled to the processor, the memory comprising instructions executable by the processor to cause the apparatus to: Receiving a reference signal (RS) set for channel measurement corresponding to a non-zero power (NZP) CSI-RS resource based on a channel state information (CSI) reporting configuration; generating a CSI report based on the channel measurement, the CSI report comprising precoding matrix information for a precoding matrix and a set of coefficients corresponding to at least one dimension of the precoding matrix; as well as The CSI report is transmitted.
2. The apparatus of claim 1 , wherein the precoding matrix comprises a first basis transform associated with a first dimension of the at least one dimension of the precoding matrix, and the set of coefficients corresponds to the precoding matrix decomposed into two subsets of coefficients, the two subsets of coefficients corresponding to different indices of the first dimension of the precoding matrix.
3. The apparatus according to claim 2, wherein a first subset of the two coefficient subsets associated with the second basis transform is associated with a second dimension of the precoding matrix, and a second subset of the two coefficient subsets associated with a third basis transform is associated with the second dimension of the precoding matrix.
4. The apparatus of claim 3, wherein the second basis transform comprises at least one of: averaging the plurality of indices of the second dimension into a smaller number of indices corresponding to the second dimension; One-to-one transformation into the same number of indices corresponding to the second dimension; selecting a subset of indices from the set of indices corresponding to the second dimension; or One of an even index or an odd index corresponding to the second dimension is selected.
5. The apparatus of claim 3, wherein the third basis transform transforms the second dimension of the precoding matrix from a domain comprising a plurality of CSI-RS ports to a domain comprising one or more spatial beams. 6 . The apparatus of claim 3 , wherein the second basis transform corresponds to the third basis transform of the same type, the third basis transform transforming the second dimension into a smaller number of indices than the second basis transform.
7. The apparatus of claim 3, wherein the third basis transform corresponds to at least one of: transforming the second dimension of the precoding matrix from a domain comprising frequency sub-bands of a plurality of channels to a domain comprising a transformed frequency domain basis; or The second dimension of the precoding matrix is transformed from a domain comprising time units of a plurality of channels to a domain comprising a transformed time domain basis.
8. A device according to claim 3, wherein the instructions are executable by the processor to cause the device to configure the third basis transform associated with the second subset of the two coefficient subsets according to a set of preconfigured transform methods, the preconfigured transform method set including at least one of a discrete Fourier transform, a discrete cosine transform, a discrete sine transform, a discrete wavelet transform, or a Haar transform.
9. The apparatus of claim 2, wherein the instructions are executable by the processor to cause the apparatus to transform the first dimension of the precoding matrix from a domain comprising multiple CSI-RS ports to a domain comprising one or more spatial beams based on the first basis transform.
10. A device according to claim 2, wherein the instructions are executable by the processor to cause the device to: represent the coefficients associated with the first subset of the two coefficient subsets with a different quantization value codebook compared to the coefficients associated with the second subset of the two coefficient subsets, and the number of bits associated with quantization and the coefficients associated with the first subset of the two coefficient subsets is greater than the number of bits associated with quantization and the coefficients associated with the second subset of the two coefficient subsets.
11. The apparatus of claim 2, wherein the instructions are executable by the processor to cause the apparatus to configure the apparatus with a higher layer parameter within the CSI reporting setting that enables or disables reporting of the first of the two coefficient subsets.
12. The apparatus of claim 1, wherein the instructions are executable by the processor to cause the apparatus to configure the apparatus to report a reporting amount corresponding to a training indicator.
13. The apparatus of claim 1, wherein the NZP CSI-RS resource comprises a plurality of CSI-RS ports.
14. A method comprising: Receiving a reference signal (RS) set for channel measurement corresponding to a non-zero power (NZP) CSI-RS resource based on a channel state information (CSI) reporting configuration; generating a CSI report based on the channel measurement, the CSI report comprising precoding matrix information for a precoding matrix and a set of coefficients corresponding to at least one dimension of the precoding matrix; as well as The CSI report is transmitted.
15. An apparatus comprising: processor; and a memory coupled to the processor, the memory comprising instructions executable by the processor to cause the apparatus to: Transmitting to a user equipment (UE) a channel state information (CSI) reporting setting for configuring the UE for at least one of channel measurement and reporting; Transmitting to the UE a reference signal (RS) set for channel measurement corresponding to a non-zero power (NZP) CSI-RS resource; as well as A CSI report is received from the UE, the CSI report including precoding matrix information for a precoding matrix and a set of coefficients corresponding to at least one dimension of the precoding matrix.
Citation Information
Cited By
Communication method, apparatus and system
WO2026189239A1
Systems and methods for model-based performance monitoring using label information
WO2026194081A1
