Realignment methods for ai-based CSI compression with historical information

Abstract

A method, network node and user equipment (UE) and apparatus for realignment methods for artificial intelligence (AI)-based channel state information (CSI) compression with historical information are disclosed. According to one aspect, a method in a UE includes receiving a configuration related to AI-based CSI compression with historical CSI using a UE part of an AI model. The method includes in response to a first indication, executing a realignment procedure of the UE part of the AI model based at least in part on the configuration.

Description

[0001] REALIGNMENT METHODS FOR AI-BASED CSI COMPRESSION WITH

[0002] HISTORICAL INFORMATION

[0003] FIELD

[0004] The present disclosure relates to wireless communications, and in particular, to realignment methods for artificial intelligence (Al)-based channel state information (CSI) compression with historical information.

[0005] BACKGROUND

[0006] The Third Generation Partnership Project (3GPP) has developed and is developing standards for Fourth Generation (4G) (also referred to as Long Term Evolution (LTE)) and Fifth Generation (5G) (also referred to as New Radio (NR)) wireless communication systems. Such systems provide, among other features, broadband communication between network nodes, such as base stations, and mobile user equipments (UE), as well as communication between network nodes and between UEs. The 3GPP is also developing standards for Sixth Generation (6G) wireless communication networks.

[0007] Artificial Intelligence (Al) and Machine Learning (ML) have been investigated, both in academia and industry, as promising tools to optimize the air interface design in wireless communication networks. One example of the AI / ML implementation for the physical layer (Al PHY) is using AI / ML for an autoencoder to improve the channel compression accuracy and / or to reduce the channel state information (CSI) feedback overhead. This topic has been discussed in 3GPP Technical Release 18 (3GPP Rel-18) and continues to be discussed in 3GPP Rel-19.

[0008] General aspects for NR 3 GPP Rel-19 AI / ML for NR air interface

[0009] In the legacy mechanism (i.e., non-AI / ML-based CSI compression), the UE reports the measured CSI-reference signals (RS) to the network node as a CSI report, based on a certain mechanism, the so-called codebook. The codebook generally defines how the UE arranges the reported bits based on (the number of) beams and taps selected by the UE to be reported to the network node and how the UE quantizes the precoding matrix.

[0010] In AI / ML-based CSI compression, an AI / ML-based autoencoder (AE) replaces at least part of the legacy mechanism. AEs may have different architectures. For example, AEs may be based on dense network nodes, multi-dimensional convolution network nodes, variational, recurrent network nodes, transformer networks, or any combination thereof. However, all AE architectures possess an encoder-bottleneck-decoder structure as illustrated in the example of FIG. 1. The codeword’s size (denoted by Y in FIG. 1) of an AE is smaller than the input data’s size (X in FIG. 1). The AE encoder thus reduces the dimensionality of the input features X down to Y. The decoder part of the AE tries to invert the encoder and reconstruct X with minimal error, according to some predefined loss function.

[0011] FIG. 2 illustrates an example of how an AE might be used for AI / ML-enhanced CSI reporting in NR. The UE measures the channel in the downlink using CSI-RS. The UE estimates that channel for each subcarrier (SC) from each base station transmit (TX) antenna and at each UE receive (RX) antenna. The estimate may be viewed as a three-dimensional channel matrix. The 3D channel matrix represents the multiple input multiple output (MIMO) channel estimated over several SCs and is input to the encoder.

[0012] The AE encoder is implemented in the UE, and the AE decoder is implemented in the network node. The output of the AE encoder is signaled from the UE to the network node over the uplink. The codeword may be viewed as a learned latent representation of the channel. Properties of the data (e.g., CSI-RS channel estimates), the channel size, uplink feedback rate, and hardware limitations of the encoder and decoder need to be considered when optimizing the AE’s architecture.

[0013] The weights and biases of an AE (with a fixed architecture) are trained to minimize the reconstruction error (the error between the input X and output X') on some training datasets. For example, the weights and biases may be trained to minimize the mean squared error, MSE ((X — X)2). To achieve good performance during live operation, the training data set should represent the actual data the AE will encounter during live operation.

[0014] In two-sided CSI compression, the output of the UE-side encoder needs to be communicated over the air interface to the network node 16 decoder with the assigned CSI reporting payload and, therefore, needs to be quantized to a finite number of bits (e.g., 1-4 bits per encoder output’s neuron) to obtain an efficient transmission as shown in the example of FIG. 3. Accordingly, a quantization layer is connected to the output of the encoder or directly included in the encoder. In an example, the quantization layer may implement scalar quantization which quantizes the output of each neuron of the encoder output layer (the bottleneck layer of AE) to generate bits to fit the CSI reporting payload in the uplink control information (UCI). Other quantization methods, e.g., vector quantization, may also be used.

[0015] Pre-processing for input data to the AE

[0016] A proper pre-processing on the input to the encoder may greatly reduce the size and complexity of designing and / or training an AI / ML model, and improve the scalability and transferability of the model. In addition, pre-processing may reduce the need for multiple models, depending on bandwidth variation and variation in the number of antenna ports at the network node, which may be a gNB. By using pre-processing, instead of directly compressing the channel (i.e., with dimensions of RX x TX x SC) as in FIG. 2, the channels are first processed to obtain another representation of the channel.

[0017] One example of the pre-processing may include transforming the channel into eigenvectors (see the example of FIG. 4). Here, the UE may perform the following steps:

[0018] 1. Compute the covariance matrix of the channel and extract the relevant eigenvectors;

[0019] 2. The covariance matrix is summed over 4 “f-units”, to get 13 “subchunks” (subbands) in frequency;

[0020] 3. For each of the 13 averaged covariance matrices, compute an eigen- decomposition and extract the 4 eigenvectors corresponding to the 4 largest eigenvalues; and

[0021] 4. Normalize the phase and magnitude of the eigenvectors.

[0022] The number of eigenvectors to feedback should be the same for all sub-chunks and depends on the rank hypothesis testing with a value between 1 and 4.

[0023] In another example, pre-processing may include transforming the channel into the beam-delay domain. The feature extraction for beam-delay reduced eigenvector-based feedback is illustrated in the example of FIG. 5. This is also called a “W2 compressor” because the feature extraction is similar to the standardized 3GPP procedures and the jargon for the matrix left to compress for the AI / ML is “W2”. The steps are as follows:

[0024] 1. The UE does a spatial domain (SD) discrete Fourier transform (DFT) on the 32x4 matrix, per resource block (RB), and selects the L strongest beams out of 16 (for one polarization). This is done in a wideband manner, including the spatial oversampling of the SD bases. Note that the same beams are used for both polarizations. The covariance of the beam-space channel is summed over, e.g., 4 RBs to produce a covariance matrix for each subband;

[0025] 2. For each covariance matrix (per subband) the UE extracts a number of eigenvectors and may select the rank, i.e. number of layers;

[0026] 3. The UE does a frequency domain DFT per layer, transforming to a delay domain, whereafter it selects the M strongest taps. The resulting tensor of dimensions 2L x number of layers x M is called the linear combination coefficients, in 3GPP jargon “W2”. The W2 matrix may be used to reconstruct the UE-suggested precoding matrices; and 4. The data is used as input in the AI / ML model. This could be the raw linear combination coefficients, or it could be enhanced with information about the selected beams and taps, noise levels, etc.

[0027] Al-based temporal-spatial-frequency (TSF) domain CSI Compression

[0028] The 3GPP Rel-18 studies on Al CSI compression focused on the compression of a channel measurement in the spatial and frequency domain. In 3 GPP Rel-19, new use cases have been considered for CSI compression to also include the temporal domain compression aspects. One category of the considered new Al CSI compression use case is to use the past CSI information to compress and reconstruct the present CSI at the UE side (encoder) and network node side (decoder) respectively. The Al model may store past CSI information from the previous slot(s) and use this information to better compress / recover the CSI of the present slot. The past information from previous slot(s) may be regarded as past CSI information, and the Al-generated CSI feedback over the air-interface for the current slot may be considered as delta CSI information on top of the past CSI information. If the channel of the current slot is correlated with the previous slot(s) and if the UE side (encoder) and network node side (decoder) have aligned past CSI information available, then, the CSI feedback overhead is expected to be further reduced as compared to the 3GPP Rel-18 CSI compression use case which considers only spatial and frequency domain CSI compression.

[0029] In the example of FIG. 6, the measured channel at time step t is used as the encoder input together with the historical CSI information at time step t-1. As mentioned in the previous section, the measured channel may also be pre-processed before it is used as the encoder input (e.g., to a form of eigenvector, W2-like format, etc.). Note that in the above example, the multiple encoder blocks may not necessarily represent a single, separate, encoder at different time steps. Depending on the model architecture the encoder at different time steps may have different parameter values, memory, states, etc. Similarly, historical CSI information may include “internal” information where the updates / information exchanges may happen inside the model itself, including the model parameter, memory, states, etc. The encoder’s output, before being transmitted to the network node, may be quantized to match the uplink / CSI report payload size.

[0030] On the network node side, the CSI report at time t is used as inputs for the decoder together with historical CSI information at time t-1. The historical CSI information on the network node side may be different from the historical CSI information on the UE side. Note that in this disclosure, the term historical information may refer to both the historical CSI information on the UE side and on the network node side. The main output of the decoder is the reconstructed channels. Note that this is only an example. The output may also be in other terms, e.g., in terms of precoder matrix.

[0031] Referring to FIG. 7, in another example, exploiting historical CSI information may also be in terms of using the same side information from the previous CSI. For example, the side information may be the Wi and / or Wf used in the pre-processing process of the previously measured channels (e.g., channels at time step 0). Information on Wi and / or Wf may be transmitted to the network node once in every cycle. For example, in the case of one cycle containing 4 CSI reports, the Wi and E / may be transmitted to the network node at time step 0 along with the (quantized) encoder outputs. At time step 1, 2, and 3, only the encoder output will be transmitted to the network node. The cycle is then repeated.

[0032] Note that the above descriptions are only examples and different architectures for exploiting the historical CSI information may also be used. In addition, a combination of the above examples may also be applied. Note also that CSI compression with historical information may be referred to as temporal-spatial-frequency (TSF) domain CSI compression.

[0033] In Al-based CSI compression with historical information, the time steps and the historical CSI information of the UE-part model (encoder) and the network node-part model (decoder) part must be aligned. The misalignment may result in performance degradation of the Al model, causing a degradation of the system performance as a whole, e.g., throughput loss. On the other hand, several circumstances may cause the time step and historical CSI information of the UE-part model and the network node-part model to misalign. Therefore, how to realign the historical CSI information of the UE part model and the network node part model from time to time is a problem.

[0034] SUMMARY

[0035] Some embodiments advantageously provide methods, network nodes and user equipments for realignment methods for artificial intelligence (Al)-based channel state information (CSI) compression with historical information.

[0036] This disclosure describes realignment mechanisms for Al-based CSI compression with historical CSI. In some embodiments, the circumstances that may trigger the realignment procedure are addressed. In some embodiments, the behavior or action that should be taken by both the UE and the network node when the realignment procedure is triggered is addressed. Some embodiments include mechanisms for realignment procedures of the UE-part model and the network node-part model for Al-based TSF CSI compression. In particular, the following are disclosed:

[0037] 1. Event / signaling that may trigger the UE-part model and network node-part model realignment procedure. The event may include:

[0038] • UE enters the discontinuous reception (DRX) inactive time;

[0039] • Receiving downlink (DL) signal of semi -persistent or aperiodic CSI report;

[0040] • Receiving explicit indication; and / or

[0041] • CSI reporting with higher rank;

[0042] 2. Behavior when the realignment procedure is triggered. The possible options may include:

[0043] • Resetting the time step and resetting the previous state and historical CSI information to be the same as the state and historical CSI at time step 0; and / or

[0044] • Use the statistic of the historical information, e.g., average, interpolation, extrapolation, using the last available state / information, etc.

[0045] In some embodiments, the alignment between the UE-part model and the network node-part model may be guaranteed. The alignment may include the alignment of the time step, historical CSI information, the internal states, internal memory, etc. Having the UE- part model and the network node-part model aligned, the potential gain of Al-based SF compression with historical CSI (i.e., TSF compression) may be well exploited, which may result in a better overall system performance, e.g., better throughput.

[0046] According to one aspect, a method in a user equipment, UE, configured to communicate with a network node includes receiving a configuration related to artificial intelligence, Al, -based channel state information, CSI, compression with historical CSI using a UE part of an Al model. The method includes, in response to a first indication, executing a realignment procedure of the UE part of the Al model based at least in part on the configuration.

[0047] According to this aspect, in some embodiments, the configuration includes one of a timer configuration and a counter configuration. In some embodiments, the first indication configures UE the to execute the realignment procedure upon expiration of a timer. In some embodiments. In some embodiments, the first indication configures the UE to execute the realignment step upon reaching a maximum time step. In some embodiments, the first indication configures the UE to execute the realignment procedure upon at least one of expiration of a drx-inactivity timer, expiration of a drx on-duration timer and start of the drx on-duration timer. In some embodiments, the first indication is included in a downlink signal. In some embodiments, the downlink signal is configured to trigger transmission of CSI reports semi-persistently or aperiodically. In some embodiments, execution of the realignment procedure occurs upon receiving a downlink signal triggering transmission of CSI reports semi-persistently or aperiodically. In some embodiments, execution of the realignment procedure occurs before transmitting a CSI report with higher rank than a rank of a previously reported CSI. In some embodiments, the first indication configures the UE to send a second indication to the network node indicating that the realignment procedure has been executed. In some embodiments, the second indication that the realignment procedure has been executed is included in the CSI report. In some embodiments, the second indication that the realignment procedure has been executed indicates a time step of historical CSI information used to produce a CSI report. In some embodiments, the realignment procedure includes resetting at least one of historical CSI information, internal states and internal memories of a UE part of an Al model. In some embodiments, the resetting includes returning a state of the UE part of the Al model to a previous initial state. In some embodiments, the resetting includes zeroing values of the historical CSI information, internal states and internal memories of a UE part of an Al model. In some embodiments, execution of the realignment procedure is based at least in part on at least one of averaging and extrapolating from historical CSI information over a plurality of previous time steps. In some embodiments, execution of the realignment procedure is based at least in part on historical CSI information used by the UE to generate a last CSI information successfully received by the network node.

[0048] According to another aspect, a UE configured to communicate with a network node is provided. The UE includes comprising processing circuitry configured to: receive a configuration related to artificial intelligence, Al, -based channel state information, CSI, compression with historical CSI using a UE part of an Al model; and in response to a first indication, execute a realignment procedure of the UE part of the Al model based at least in part on the configuration.

[0049] According to this aspect, in some embodiments, the configuration includes one of a timer configuration and a counter configuration. In some embodiments, the first indication configures the UE to execute the realignment procedure upon expiration of a timer. In some embodiments, the first indication configures the UE to execute the realignment step upon reaching a maximum time step. In some embodiments, the first indication configures the UE to execute the realignment procedure upon at least one of expiration of a drx-inactivity timer, expiration of a drx on-duration timer and start of the drx on-duration timer. In some embodiments, the first indication is included in a downlink signal. In some embodiments, the downlink signal is configured to trigger transmission of CSI reports semi-persistently or aperiodically. In some embodiments, execution of the realignment procedure occurs upon receiving a downlink signal triggering transmission of CSI reports semi-persistently or aperiodically. In some embodiments, execution of the realignment procedure occurs before transmitting a CSI report with higher rank than a rank of a previously reported CSI. In some embodiments, the first indication configures the UE to send a second indication to the network node indicating that the realignment procedure has been executed. In some embodiments, the second indication that the realignment procedure has been executed is included in the CSI report. In some embodiments, the second indication that the realignment procedure has been executed indicates a time step of historical CSI information used to produce a CSI report. In some embodiments, the realignment procedure includes resetting at least one of historical CSI information, internal states and internal memories of a UE part of an Al model. In some embodiments, the resetting includes returning a state of the UE part of the Al model to a previous initial state. In some embodiments, the resetting includes zeroing values of the historical CSI information, internal states and internal memories of a UE part of an Al model. In some embodiments, execution of the realignment procedure is based at least in part on at least one of averaging and extrapolating from historical CSI information over a plurality of previous time steps. In some embodiments, execution of the realignment procedure is based at least in part on historical CSI information used by the UE to generate a last CSI information successfully received by the network node.

[0050] According to yet another aspect, a method implemented in a network node that is configured to communicate with a user equipment, UE, is provided. The method includes transmitting a configuration related to artificial intelligence, Al, -based channel state information, CSI, compression with historical CSI. The method also includes receiving a CSI report from the UE. The method further includes executing a realignment procedure of a network node part of an Al model, based at least in part on the CSI report.

[0051] According to the aspect, in some embodiments, the CSI report is based at least in part on a corresponding realignment procedure of a UE part of the Al model. In some embodiments, the configuration configures the UE to execute the corresponding realignment procedure upon expiration of a timer. In some embodiments, the CSI report includes an indication that the corresponding realignment procedure has been executed. In some embodiments, the configuration configures the UE to execute the corresponding realignment procedure upon reaching a maximum time step. In some embodiments, the configuration configures the UE to execute the corresponding realignment procedure upon at least one of expiration of a drx-inactivity timer, expiration of a drx on-duration timer and start of the drx on-duration timer. In some embodiments, the CSI report includes a time step of historical CSI information used to produce the CSI report.

[0052] According to yet another aspect, a network node configured to communicate with a user equipment, UE, the network node comprising processing circuitry configured to: transmit a configuration related to artificial intelligence, Al, -based channel state information, CSI, compression with historical CSI; receive a CSI report from the UE; and execute a realignment procedure of a network node part of an Al model, based at least in part on the CSI report.

[0053] According to this aspect, in some embodiments, the CSI report is based at least in part on a corresponding realignment procedure of a UE part of the Al model. In some embodiments, the configuration configures the UE to execute the corresponding realignment procedure upon expiration of a timer. In some embodiments, the CSI report includes an indication that the corresponding realignment procedure has been executed. In some embodiments, the configuration configures the UE to execute the corresponding realignment procedure upon reaching a maximum time step. In some embodiments, the configuration configures the UE to execute the corresponding realignment procedure upon at least one of expiration of a drx-inactivity timer, expiration of a drx on-duration timer and start of the drx on-duration timer. In some embodiments, the CSI report includes a time step of historical CSI information used to produce the CSI report.

[0054] BRIEF DESCRIPTION OF THE DRAWINGS

[0055] A more complete understanding of the present embodiments, and the attendant advantages and features thereof, will be more readily understood by reference to the following detailed description when considered in conjunction with the accompanying drawings wherein:

[0056] FIG. 1 is an illustration of a fully connected autoencoder;

[0057] FIG. 2 illustrates using an autoencoder for CSI compression;

[0058] FIG. 3 illustrates a quantization operation at the output of an encoder;

[0059] FIG. 4 shows preprocessing of the channel into eigenvectors;

[0060] FIG. 5 shows preprocessing of the channel into a beam delay domain;

[0061] FIG. 6 is an example of CSI compression with historical CSI; FIG. 7 is another example of CSI compression with historical CSI;

[0062] FIG. 8 is a schematic diagram of an example network architecture illustrating a communication system according to principles disclosed herein;

[0063] FIG. 9 is a block diagram of a network node in communication with a user equipment over a wireless connection according to some embodiments of the present disclosure;

[0064] FIG. 10 is a flowchart of an example process in a network node for realignment methods for artificial intelligence (Al)-based channel state information (CSI) compression with historical information according to some embodiments of the present disclosure; and

[0065] FIG. 11 is a flowchart of an example process in a user equipment for realignment methods for artificial intelligence (Al)-based channel state information (CSI) compression with historical information according to some embodiments of the present disclosure;

[0066] FIG. 12 is a flowchart of another example process in a network node for realignment methods for artificial intelligence (Al)-based channel state information (CSI) compression with historical information according to some embodiments of the present disclosure;

[0067] FIG. 13 is a flowchart of another example process in a user equipment for realignment methods for artificial intelligence (Al)-based channel state information (CSI) compression with historical information according to some embodiments of the present disclosure; and

[0068] FIG. 14 is an example of resetting of historical CSI information when an inactivity timer expires.

[0069] DETAILED DESCRIPTION

[0070] Before describing in detail example embodiments, it is noted that the embodiments reside primarily in combinations of apparatus components and processing steps related to realignment methods for artificial intelligence (Al)-based channel state information (CSI) compression with historical information. Accordingly, components have been represented where appropriate by conventional symbols in the drawings, showing only those specific details that are pertinent to understanding the embodiments so as not to obscure the disclosure with details that will be readily apparent to those of ordinary skill in the art having the benefit of the description herein.

[0071] As used herein, relational terms, such as “first” and “second,” “top” and “bottom,” and the like, may be used solely to distinguish one entity or element from another entity or element without necessarily requiring or implying any physical or logical relationship or order between such entities or elements. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the concepts described herein. As used herein, the singular forms “a”, “an” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms “comprises,” “comprising,” “includes” and / or “including” when used herein, specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0072] In embodiments described herein, the joining term, “in communication with” and the like, may be used to indicate electrical or data communication, which may be accomplished by physical contact, induction, electromagnetic radiation, radio signaling, infrared signaling or optical signaling, for example. One having ordinary skill in the art will appreciate that multiple components may interoperate and modifications and variations are possible of achieving the electrical and data communication.

[0073] In some embodiments described herein, the term “coupled,” “connected,” and the like, may be used herein to indicate a connection, although not necessarily directly, and may include wired and / or wireless connections.

[0074] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the concepts described herein. As used herein, the singular forms “a”, “an” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms “comprises,” “comprising,” “includes” and / or “including” when used herein, specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0075] The term “network node” used herein may be any kind of network node comprised in a radio network which may further comprise any of base station (BS), radio base station, base transceiver station (BTS), base station controller (BSC), radio network controller (RNC), g Node B (gNB), evolved Node B (eNB or eNodeB), Node B, multi -standard radio (MSR) radio node such as MSR BS, multi-cell / multicast coordination entity (MCE), relay node, donor node controlling relay, radio access point (AP), transmission points, transmission nodes, Remote Radio Unit (RRU) Remote Radio Head (RRH), a core network node (e.g., mobile management entity (MME), self-organizing network (SON) node, a coordinating node, positioning node, MDT node, etc.), an external node (e.g., 3rd party node, a node external to the current network), nodes in distributed antenna system (DAS), a spectrum access system (SAS) node, an element management system (EMS), etc. The network node may also comprise test equipment. The term “radio node” used herein may be used to also denote a user equipment (UE) such as a wireless device (WD) or a radio network node.

[0076] In some embodiments, the non-limiting terms wireless device (WD) or a user equipment (UE) are used interchangeably. The UE herein may be any type of user equipment capable of communicating with a network node or another UE over radio signals, such as a wireless device (WD). The UE may also be a radio communication device, target device, device to device (D2D) UE, machine type UE or UE capable of machine to machine communication (M2M), low-cost and / or low-complexity UE, a sensor equipped with UE, Tablet, mobile terminals, smart phone, laptop embedded equipped (LEE), laptop mounted equipment (LME), USB dongles, Customer Premises Equipment (CPE), an Internet of Things (loT) device, or a Narrowband loT (NB-IOT) device etc.

[0077] Also, in some embodiments the generic term “radio network node” is used. It may be any kind of a radio network node which may comprise any of base station, radio base station, base transceiver station, base station controller, network controller, RNC, evolved Node B (eNB), Node B, gNB, Multi -cell / multicast Coordination Entity (MCE), relay node, access point, radio access point, Remote Radio Unit (RRU) Remote Radio Head (RRH).

[0078] Note that although terminology from one particular wireless system, such as, for example, 3GPP LTE and / or New Radio (NR), may be used in this disclosure, this should not be seen as limiting the scope of the disclosure to only the aforementioned system. Other wireless systems, including without limitation Wide Band Code Division Multiple Access (WCDMA), Worldwide Interoperability for Microwave Access (WiMax), Ultra Mobile Broadband (UMB) and Global System for Mobile Communications (GSM), may also benefit from exploiting the ideas covered within this disclosure.

[0079] Note further, that functions described herein as being performed by a user equipment or a network node may be distributed over a plurality of user equipments and / or network nodes. In other words, it is contemplated that the functions of the network node and user equipment described herein are not limited to performance by a single physical device and, in fact, may be distributed among several physical devices.

[0080] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs. It will be further understood that terms used herein should be interpreted as having a meaning that is consistent with their meaning in the context of this specification and the relevant art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.

[0081] Some embodiments are directed to realignment methods for artificial intelligence (Al)-based channel state information (CSI) compression with historical information.

[0082] Returning to the drawing figures, in which like elements are referred to by like reference numerals, there is shown in FIG. 8 a schematic diagram of a communication system 10, according to an embodiment, such as a 3 GPP -type cellular network that may support standards such as LTE and / or NR (5G), which comprises an access network 12, such as a radio access network, and a core network 14. The access network 12 comprises a plurality of network nodes 16a, 16b, 16c (referred to collectively as network nodes 16), such as NBs, eNBs, gNBs or other types of wireless access points, each defining a corresponding coverage area 18a, 18b, 18c (referred to collectively as coverage areas 18). Each network node 16a, 16b, 16c is connectable to the core network 14 over a wired or wireless connection 20. A first user equipment (UE) 22a located in coverage area 18a is configured to wirelessly connect to, or be paged by, the corresponding network node 16a. A second UE 22b in coverage area 18b is wirelessly connectable to the corresponding network node 16b. While a plurality of UEs 22a, 22b (collectively referred to as user equipments 22) are illustrated in this example, the disclosed embodiments are equally applicable to a situation where a sole UE is in the coverage area or where a sole UE is connecting to the corresponding network node 16. Note that although only two UEs 22 and three network nodes 16 are shown for convenience, the communication system may include many more UEs 22 and network nodes 16.

[0083] Also, it is contemplated that a UE 22 may be in simultaneous communication and / or configured to separately communicate with more than one network node 16 and more than one type of network node 16. For example, a UE 22 may have dual connectivity with a network node 16 that supports LTE and the same or a different network node 16 that supports NR. As an example, UE 22 may be in communication with an eNB for LTEZE- UTRAN and a gNB for NR / NG-RAN.

[0084] A network node 16 (eNB or gNB) is configured to include an AE decoder unit 24 which is configured to execute a realignment procedure of a network node part of an Al model, based at least in part on a CSI report received from the UE. A user equipment 22 is configured to include an AE encoder unit 26 which is configured to, in response to a first indication, execute the realignment procedure of a UE part of an Al mode based at least in part on the configuration. Example implementations, in accordance with an embodiment, of the UE 22 and network node 16 discussed in the preceding paragraphs will now be described with reference to FIG. 9.

[0085] The communication system 10 includes a network node 16 provided in a communication system 10 and including hardware 28 enabling it to communicate with the UE 22. The hardware 28 may include a radio interface 30 for setting up and maintaining at least a wireless connection 32 with a UE 22 located in a coverage area 18 served by the network node 16. The radio interface 30 may be formed as or may include, for example, one or more RF transmitters, one or more RF receivers, and / or one or more RF transceivers. The radio interface 30 includes an array of antennas 34 to radiate and receive signal(s) carrying electromagnetic waves.

[0086] In the embodiment shown, the hardware 28 of the network node 16 further includes processing circuitry 36. The processing circuitry 36 may include a processor 38 and a memory 40. In particular, in addition to or instead of a processor, such as a central processing unit, and memory, the processing circuitry 36 may comprise integrated circuitry for processing and / or control, e.g., one or more processors and / or processor cores and / or FPGAs (Field Programmable Gate Array) and / or ASICs (Application Specific Integrated Circuitry) adapted to execute instructions. The processor 38 may be configured to access (e.g., write to and / or read from) the memory 40, which may comprise any kind of volatile and / or nonvolatile memory, e.g., cache and / or buffer memory and / or RAM (Random Access Memory) and / or ROM (Read-Only Memory) and / or optical memory and / or EPROM (Erasable Programmable Read-Only Memory).

[0087] Thus, the network node 16 further has software 42 stored internally in, for example, memory 40, or stored in external memory (e.g., database, storage array, network storage device, etc.) accessible by the network node 16 via an external connection. The software 42 may be executable by the processing circuitry 36. The processing circuitry 36 may be configured to control any of the methods and / or processes described herein and / or to cause such methods, and / or processes to be performed, e.g., by network node 16. Processor 38 corresponds to one or more processors 38 for performing network node 16 functions described herein. The memory 40 is configured to store data, programmatic software code and / or other information described herein. In some embodiments, the software 42 may include instructions that, when executed by the processor 38 and / or processing circuitry 36, causes the processor 38 and / or processing circuitry 36 to perform the processes described herein with respect to network node 16. For example, processing circuitry 36 of the network node 16 may include an AE decoder unit 24 which is configured to execute a realignment procedure of a network node part of an Al model, based at least in part on a C SI report received from the UE.

[0088] The communication system 10 further includes the UE 22 already referred to. The UE 22 may have hardware 44 that may include a radio interface 46 configured to set up and maintain a wireless connection 32 with a network node 16 serving a coverage area 18 in which the UE 22 is currently located. The radio interface 46 may be formed as or may include, for example, one or more RF transmitters, one or more RF receivers, and / or one or more RF transceivers. The radio interface 46 includes an array of antennas 48 to radiate and receive signal(s) carrying electromagnetic waves.

[0089] The hardware 44 of the UE 22 further includes processing circuitry 50. The processing circuitry 50 may include a processor 52 and memory 54. In particular, in addition to or instead of a processor, such as a central processing unit, and memory, the processing circuitry 50 may comprise integrated circuitry for processing and / or control, e.g., one or more processors and / or processor cores and / or FPGAs (Field Programmable Gate Array) and / or ASICs (Application Specific Integrated Circuitry) adapted to execute instructions. The processor 52 may be configured to access (e.g., write to and / or read from) memory 54, which may comprise any kind of volatile and / or nonvolatile memory, e.g., cache and / or buffer memory and / or RAM (Random Access Memory) and / or ROM (Read-Only Memory) and / or optical memory and / or EPROM (Erasable Programmable Read-Only Memory).

[0090] Thus, the UE 22 may further comprise software 56, which is stored in, for example, memory 54 at the UE 22, or stored in external memory (e.g., database, storage array, network storage device, etc.) accessible by the UE 22. The software 56 may be executable by the processing circuitry 50. The software 56 may include a client application 58. The client application 58 may be operable to provide a service to a human or non-human user via the UE 22.

[0091] The processing circuitry 50 may be configured to control any of the methods and / or processes described herein and / or to cause such methods, and / or processes to be performed, e.g., by UE 22. The processor 52 corresponds to one or more processors 52 for performing UE 22 functions described herein. The UE 22 includes memory 54 that is configured to store data, programmatic software code and / or other information described herein. In some embodiments, the software 56 and / or the client application 58 may include instructions that, when executed by the processor 52 and / or processing circuitry 50, causes the processor 52 and / or processing circuitry 50 to perform the processes described herein with respect to UE 22. For example, the processing circuitry 50 of the user equipment 22 may include an AE encoder unit 26 which is configured to, in response to a first indication, execute the realignment procedure of a UE part of an Al mode based at least in part on the configuration.

[0092] In some embodiments, the inner workings of the network node 16 and UE 22 may be as shown in FIG. 9 and independently, the surrounding network topology may be that of FIG. 8.

[0093] The wireless connection 32 between the UE 22 and the network node 16 is in accordance with the teachings of the embodiments described throughout this disclosure. More precisely, the teachings of some of these embodiments may improve the data rate, latency, and / or power consumption and thereby provide benefits such as reduced user waiting time, relaxed restriction on file size, better responsiveness, extended battery lifetime, etc. In some embodiments, a measurement procedure may be provided for the purpose of monitoring data rate, latency and other factors on which the one or more embodiments improve.

[0094] Although FIGS. 8 and 9 show various “units” such as Al configuration unit 24 and realignment unit 26 as being within a respective processor, it is contemplated that these units may be implemented such that a portion of the unit is stored in a corresponding memory within the processing circuitry. In other words, the units may be implemented in hardware or in a combination of hardware and software within the processing circuitry.

[0095] FIG. 10 is a flowchart of an example process in a network node 16 configured according to principles disclosed herein. One or more blocks described herein may be performed by one or more elements of network node 16 such as by one or more of processing circuitry 36 (including the Al configuration unit 24), processor 38, and / or radio interface 30. Network node 16 such as via processing circuitry 36 and / or processor 38 and / or radio interface 30 is configured to transmit a configuration related to artificial intelligence (AI)- based channel state information (CSI) compression with historical CSI (Block S10). The process also includes transmitting an indication to execute a realignment procedure based at least in part on the configuration (Block S12).

[0096] In some embodiments, the process includes receiving an indication that the realignment procedure has been executed. In some embodiments, the configuration includes one of a timer and a counter. In some embodiments, the indication to execute the realignment procedure includes a timer to expire prior to execution of the realignment procedure. In some embodiments, the indication to execute the realignment procedure includes a maximum step time to execution of the realignment procedure. In some embodiments, the indication to execute the realignment procedure includes a drx-inactivity timer and a drx on-duration timer. In some embodiments, the method includes transmitting a downlink signal triggering transmission of CSI reports semi -persistently or aperiodically. In some embodiments, the indication to execute the realignment procedure causes the UE to transmit a CSI report with higher rank that a previously reported CSI.

[0097] FIG. 11 is a flowchart of an example process in a user equipment 22 according to some embodiments of the present disclosure. One or more blocks described herein may be performed by one or more elements of user equipment 22 such as by one or more of processing circuitry 50 (including the realignment unit 26), processor 52, and / or radio interface 46. User equipment 22 such as via processing circuitry 50 and / or processor 52 and / or radio interface 46 is configured to receive a configuration related to artificial intelligence (Al)-based channel state information (CSI) compression with historical CSI (Block S14). The process includes receiving an indication to execute a realignment procedure (Block SI 6). The process also includes executing the realignment procedure based at least in part on the configuration (Block SI 8).

[0098] In some embodiments, the method includes indicating to the network node that the realignment procedure has been executed. In some embodiments, the configuration includes one of a timer and a counter. In some embodiments, the indication to execute the realignment procedure includes a timer to expire prior to execution of the realignment procedure. In some embodiments, the indication to execute the realignment procedure includes a maximum step time to execution of the realignment procedure. In some embodiments, the indication to execute the realignment procedure includes a drx-inactivity timer and a drx on-duration timer. In some embodiments, execution of the realignment procedure occurs upon receiving a downlink signal triggering transmission of CSI reports semi-persistently or aperiodically. In some embodiments, execution of the realignment procedure occurs before transmitting a CSI report with higher rank that a previously reported CSI.

[0099] FIG. 12 is a flowchart of an example process in a network node 16 configured according to principles disclosed herein. One or more blocks described herein may be performed by one or more elements of network node 16 such as by one or more of processing circuitry 36 (including the Al configuration unit 24), processor 38, and / or radio interface 30. Network node 16 such as via processing circuitry 36 and / or processor 38 and / or radio interface 30 is configured to transmit a configuration related to artificial intelligence, AI,- based channel state information, CSI, compression with historical CSI (Block S20). The method also includes receiving a CSI report from the UE (Block S22). The method further includes executing a realignment procedure of a network node part of an Al model, based at least in part on the CSI report (Block S24).

[0100] According to the aspect, in some embodiments, the CSI report is based at least in part on a corresponding realignment procedure of a UE part of the Al model. In some embodiments, the configuration configures the UE to execute the corresponding realignment procedure upon expiration of a timer. In some embodiments, the CSI report includes an indication that the corresponding realignment procedure has been executed. In some embodiments, the configuration configures the UE to execute the corresponding realignment procedure upon reaching a maximum time step. In some embodiments, the configuration configures the UE to execute the corresponding realignment procedure upon at least one of expiration of a drx-inactivity timer, expiration of a drx on-duration timer and start of the drx on-duration timer. In some embodiments, the CSI report includes a time step of historical CSI information used to produce the CSI report.

[0101] FIG. 13 is a flowchart of an example process in a user equipment 22 according to some embodiments of the present disclosure. One or more blocks described herein may be performed by one or more elements of user equipment 22 such as by one or more of processing circuitry 50 (including the realignment unit 26), processor 52, and / or radio interface 46. User equipment 22 such as via processing circuitry 50 and / or processor 52 and / or radio interface 46 is configured to receive a configuration related to artificial intelligence, Al, -based channel state information, CSI, compression with historical CSI using a UE part of an Al model (Block S26). The method includes, in response to a first indication, executing a realignment procedure of the UE part of the Al model based at least in part on the configuration (Block S28).

[0102] According to this aspect, in some embodiments, the configuration includes one of a timer configuration and a counter configuration. In some embodiments, the first indication configures UE the to execute the realignment procedure upon expiration of a timer. In some embodiments. In some embodiments, the first indication configures the UE to execute the realignment step upon reaching a maximum time step. In some embodiments, the first indication configures the UE to execute the realignment procedure upon at least one of expiration of a drx-inactivity timer, expiration of a drx on-duration timer and start of the drx on-duration timer. In some embodiments, the first indication is included in a downlink signal. In some embodiments, the downlink signal is configured to trigger transmission of CSI reports semi-persistently or aperiodically. In some embodiments, execution of the realignment procedure occurs upon receiving a downlink signal triggering transmission of CSI reports semi-persistently or aperiodically. In some embodiments, execution of the realignment procedure occurs before transmitting a CSI report with higher rank than a rank of a previously reported CSI. In some embodiments, the first indication configures the UE to send a second indication to the network node indicating that the realignment procedure has been executed. In some embodiments, the second indication that the realignment procedure has been executed is included in the CSI report. In some embodiments, the second indication that the realignment procedure has been executed indicates a time step of historical CSI information used to produce a CSI report. In some embodiments, the realignment procedure includes resetting at least one of historical CSI information, internal states and internal memories of a UE part of an Al model. In some embodiments, the resetting includes returning a state of the UE part of the Al model to a previous initial state. In some embodiments, the resetting includes zeroing values of the historical CSI information, internal states and internal memories of a UE part of an Al model. In some embodiments, execution of the realignment procedure is based at least in part on at least one of averaging and extrapolating from historical CSI information over a plurality of previous time steps. In some embodiments, execution of the realignment procedure is based at least in part on historical CSI information used by the UE to generate a last CSI information successfully received by the network node.

[0103] Having described the general process flow of arrangements of the disclosure and having provided examples of hardware and software arrangements for implementing the processes and functions of the disclosure, the sections below provide details and examples of arrangements for realignment methods for artificial intelligence (Al)-based channel state information (CSI) compression with historical information.

[0104] As described herein, the concept of “network (network node)” and / or a gNB may be understood as a generic network node, gNB, base station, unit within the base station, relay node, core network node, a core network node, or a device supporting D2D communication. The node may be deployed in a 3GPP 5G network, or a 3GPP 6G network. Moreover, although the term AI / ML model uses a single form, it should be well understood that it should not prevent the implementation of more than one AI / ML model. The UE may either be configured or autonomously switch between the models depending on certain conditions and / or proprietary implementations. In addition, the term historical CSI information may also be a general term which may include historical CSI information which is generated during CSI report generation, internal memories and internal states of the model, etc. Configuration on the realignment procedure.

[0105] In some embodiments, the alignment between the UE-part model (encoder) on the UE side and the network node-part model (decoder) on the network node side may be done periodically. For example, the UE 22 and the network node 16 may reset the historical CSI information in every N time steps.

[0106] In some embodiments, the configuration of the Al-based CSI compression (e.g., in the radio resource control (RRC) configuration) may include a parameter that determines the number of time steps (and thus, the number of CSI reports) before the UE 22 resets the historical CSI information. The parameter may be called, e.g., max time steps where the value may be an integer value. The time step may be restarted (e.g., restart from 0 or 1) once the resetting procedure takes place.

[0107] In some embodiments, the configuration may be in terms of a timer. For example, a parameter of resetting timer may be introduced. The UE 22 may reset the historical CSI information once the resetting timer expires. The resetting timer may then also be restarted.

[0108] Historical CSI information is useful when there is a correlation between the previous CSI and the current CSI. When the time between two consecutive CSI reports is long, it may be better for the UE 22 (and the network node 16) to reset the historical CSI information. In some embodiments, a parameter, e.g., model inactivity timer may be introduced. The value of the timer may be in terms of milli-seconds (ms), symbols, or slots, for example.

[0109] In some embodiments, the timer starts when a CSI report is generated by the encoder. The UE 22 may restart the timer each time the UE 22 generates a CSI report using its encoder.

[0110] In some embodiments, the timer starts when a CSI report generated by the encoder is transmitted by the UE 22 to the network node 16. In this embodiment, the UE 22 may restart the timer each time the UE 22 reports a CSI report generated by the encoder.

[0111] If the timer expires (e.g., if the distance between two consecutive CSIs is too far), the UE 22 may reset the historical CSI information.

[0112] Note that since the inactivity timer is used to reset the historical CSI information when it expires, the inactivity timer may be thought of as an inactivity timer for CSI reporting. That is, the timer expires e.g., when there have not been any CSI reports transmitted for a predefined or preconfigured duration of the timer and the expiry of the timer results in historical CSI information being reset at the UE 22. In the same slot the timer expires, the network node 16 may also reset its historical CSI information at the network node side. Note that resetting the historical CSI information at the UE 22 and the network node 16 is an example of properties of the UE-part model on the UE side and the network node-part model on the network node 16 (or gNB) side being realigned.

[0113] An example illustrating the resetting of historical CSI information when the inactivity timer expires is illustrated in FIG. 14.

[0114] • At slot n0, the UE 22 receives a control message (e.g., either a downlink control information (DCI) or medium access control (MAC) control element (CE)) that triggers or activates CSI reporting;

[0115] • The UE 22 transmits CSI report at slot at which time the inactivity timer is started by the UE 22. Note that the network node 16 may also start the inactivity timer at slot n^,

[0116] • In FIG. 14, there is another CSI report transmitted by the UE 22 at slot n2. At slot n2, the inactivity timer is (re)started by the UE 22. The network node 16 may also (re)start the inactivity timer at slot n2; and / or

[0117] • At slot n3, the inactivity timer expires, and the UE 22 resets the historical CSI information at the encoder. Similarly, the network node 16 may also reset the historical CSI information at the decoder.

[0118] Some embodiments described above disclose that the inactivity timer starts when a CSI report is transmitted by the UE 22. In some embodiments, the inactivity timer may start when the UE 22 receives a control message (e.g., either a DCI or MAC CE) for triggering or activating CSI reporting.

[0119] The maximum value of model inactivity timer may depend on the UE capability, i.e., the UE 22 may report to the network (as part of UE capability reporting) the maximum value of model inactivity timer that may be supported by the UE 22. Note that in some embodiments, the maximum value may be classified for a plurality of UE speeds. For example, the UE 22 may report (in its UE capability report) a first maximum value of model inactivity timer for a first (maximum) UE speed and a second maximum value of model inactivity timer for a second (maximum) UE speed.

[0120] Indication that may trigger realignment procedure.

[0121] Entering the DRX inactive time

[0122] When the UE 22 is configured with discontinuous reception (DRX), the UE 22 may be in the active or inactive time. The UE 22 is in the active time, e.g., when the UE 22 is in the drx on-duration (i.e., the drx on-duration timer is not expired) or when the drx inactivity timer is not expired. If the drx-inactivity timer expires, and the UE 22 is not yet in the drx on-duration, the UE 22 is in the inactive state. During an inactive state, the UE 22 is not expected to transmit the CSI report. This may hinder the UE 22 (and the network node 16) from using the historical CSI information as the input for the encoder (and the decoder). First, the time between two drx on-durations may be long and thus, the available historical CSI information may already be outdated. Second, the UE 22 may not have information on a certain time step. For example, the UE 22 may be configured with a drx cycle of 20ms of which 10ms is the drx on-duration and is configured with CSI-RS and CSI report periodicity of 5ms. The first two CSI reports (say the first and the second) may fall inside the drx on- duration while the other two (say the third and fourth) may fall outside the drx on-duration. Here, the CSI report at time step 0 may not have historical CSI information while the CSI report at time step 1 may have historical CSI information from time step 0. As the UE 22 does not generate a CSI report for time step 2 and time step 3, the CSI report at time step 4 may have no proper historical CSI input (the historical CSI information from time step 2 and time step 3 are missing). This may cause a performance degradation on the CSI report, especially when two entities (the UE 22 and the network node 16) take different approaches to handling such a situation.

[0123] In some embodiments, the change in the DRX state may serve as an indication to execute the realignment procedure (e.g., resetting the historical CSI information). In some embodiments, a transition from the drx inactive state to the drx active state may serve as an implicit indication to conduct the realignment procedure. For example, the start of the drx on-duration timer may serve as an implicit indication to conduct the realignment procedure. In some embodiments, transition from the drx active state to the drx inactive state may serve as an implicit indication to conduct the realignment procedure. Such indication may, for example, include the end of drx on-duration timer or the end of the drx inactivity timer.

[0124] Receiving a DL signal to transmit a semi persistent or aperiodic CSI report.

[0125] A UE 22 may be configured with at least one of the periodic, semi-persistent, and aperiodic CSI reports. When configured with periodic CSI reporting, the UE 22 may transmit the CSI report periodically starting from the time the configuration is received by the UE 22 without any indication of when the reporting ends. For the case of the semi- persistent CSI report, there will be an indication to start multiple CSI report occasions. For example, a single indication may represent an indication to transmit N CSI reports. After the UE 22 transmits N periodic CSI reports, the UE 22 may stop transmitting the CSI reports until a further indication. For the case of aperiodic CSI reporting, the UE 22 may transmit a single CSI report once the UE 22 is indicated by the network node 16. In particular, for the case of semi persistent and aperiodic CSI reporting, there may be some gap between two consecutive CSI reports, e.g., the gap between two aperiodic CSI reports, or the gap between the first CSI report of the current semi persistent CSI report and the last CSI report of the previous semi persistent CSI report.

[0126] In some embodiments, an indication of the semi-persistent or aperiodic CSI report may serve as an implicit indication to conduct a realignment procedure. In some embodiments, the realignment procedure may instead be executed by the UE 22 (and the network node 16) once the last CSI report within a semi-persistent CSI report is generated / transmitted by the UE 22 (or received by the network node 16). In another example, for the case of the aperiodic CSI report, the UE 22 may conduct the realignment procedure each time the UE 22 finishes generating and / or transmitting the CSI report.

[0127] In some embodiments, whether the UE 22 conducts the realignment procedure may depend on the time between two indications. For example, for the case of an aperiodic CSI report, if the time between a CSI report indication and the previous CSI report indication is larger than a certain threshold, the UE 22 may conduct the realignment procedure for the current CSI report calculation. In some embodiments, a timer may be utilized, where the timer represents the maximum time / gap between two consecutive CSI reports (time step) for which the historical CSI is still useful for the encoder (and the decoder). In some embodiments, the timer starts (and restarts) each time an aperiodic CSI report indication is received by the UE 22. In some embodiments, the timer may start and restart each time an aperiodic CSI report is generated or transmitted to the network node 16 by the UE 22. For the case of a semi -persistent CSI report, the timer may represent the maximum time / gap between the first CSI report of a semi-persistent CSI report and the last CSI report of the previous semi -persistent CSI report. Here, the timer may start and restart each time a CSI report is generated or transmitted to the network node 16.

[0128] Receiving an explicit realignment indication from the network node.

[0129] In some embodiment, an explicit indication may be used. For example, this may resolve the problem of a missing CSI report from the UE 22 to the network node 16. Note that this is an example implementation of an explicit realignment indication received by the UE 22 from the network node 16. The reason behind the explicit indication may depend on the network node implementation.

[0130] In some embodiments, a bitfield in the downlink control information (DCI) or medium access control (MAC) control element (CE) may include an indication to the UE 22 to take the realignment procedure (e.g., resetting the historical CSI information). For example, the bitfield may be included in the indication for the aperiodic or semi-persistent CSI report. In some embodiments, the bitfield may have a size of 1 bit, e.g., where 0 may represent an indication to not reset the historical CSI information while 1 may represent an indication to reset the historical CSI information. In some embodiments, the bitfield may have a size larger than 1 bit, e.g., 2 bits. The value of the bitfield may further indicate how many steps the UE 22 needs to go back, or which time step the UE 22 may use to generate the CSI report. For example, the values of 00, 01, 10, and 11 may represent an indication to return to time step 0, to return to a first value of time step, to return to a second value of time step, and to not conduct realignment procedure, respectively. The values (the first value and the second value) in this example may be configured by the network node 16 or may be predetermined in the standard text.

[0131] In some embodiments, the network node 16 may explicitly signal the realignment of a subset of historical CSI information. For example, where the historical CSI information includes spatial domain basis vectors (Wl past) and frequency domain basis vectors (W-past) of a past channel measurement reported at a previous time slot t0, the network node 16 may explicitly signal to realign one or both of {Wl past, W-pastj. Accordingly, the network node 16 may indicate a bitfield of size 2, where the bitfield may refer to the following: a) 00: no realignment, b) 01 : realign Wl past, c) 10: realign Wypast, and d) 11 : realign both { l,past' ,past}-

[0132] Transmitting CSI report with a higher rank

[0133] In some embodiments, the UE 22 may also initiate the realignment procedure. For example, it is possible that the current link quality is better than the previous link quality, and thus, the UE 22 would prefer a downlink (DL) transmission with a higher rank compared to the previous DL transmission. To obtain the higher rank transmission, the UE 22 may transmit a CSI report containing the report with a higher rank compared to the previous report. For example, the UE 22 may report CSI for rank = 2 in the previous CSI report occasion and may need to report CSI for rank = 4 in the current CSI report occasion. The UE 22 (and the network node 16), however, may not have historical CSI information for layer 3 and layer 4.

[0134] In some embodiments, reporting with a higher rank than the rank of the previous DL transmission may serve as an (implicit) indication from the UE 22 to the network node 16 that a realignment procedure has taken place on the UE side. Accordingly, the network node 16 may execute the realignment procedure in the network node-part model (decoder). In some embodiments, the realignment procedure may be done for all layers, including the layers from the previously reported rank. This, for example, is suitable for the case of the Al model that processes the input jointly for all layers to produce the CSI report for the respected rank.

[0135] In some embodiments, the realignment procedure may be done on a per-layer basis. For example, in time steps 0, 1, and 2 the UE 22 may report CSI with rank = 4, rank = 2, and rank = 4, respectively. For a CSI report in time step 2, the realignment procedure may be done for layer 3 and layer 4 while the CSI report for layer 1 and layer 2 may be generated without any realignment procedures. Note that whether the realignment procedure needs to be taken for layer 3 and layer 4 may further consider the distance between the CSI report for time step 3 and time step 2 (for the above example). If the distance is above a predetermined or configured threshold, the realignment procedure may be taken, while if the distance is equal to or below the configured or predetermined threshold, the realignment procedure may be omitted.

[0136] Experiencing channel that may cause performance degradation for the subsequent CSI report.

[0137] The compression quality of the Al-based CSI report with historical CSI information depends on the quality of the previous outputs of the model. The quality of the output itself depends on, e.g., how easily the channel may be compressed by the model. In other words, the channel condition at time step t may affect the quality of the CSI report at time step t” where f > t. Therefore, the UE 22 may omit a certain historical CSI information either by resetting to time step 0 or by using other alternatives as the historical CSI information, e.g., by using the historical CSI information at f where t- f > 1. To indicate that a realignment procedure has been taken by the UE 22, the UE 22 may need to indicate such an operation to the network node 16. In some embodiments, such an indication may be embedded in the CSI report. Note that a CSI report may contain more than 1 part, e.g., 2 parts, where Part 1 may contain information on the number of reported ranks, the selected beams and taps, etc., while Part 2 may contain precoder matrix information (PMI) information. In some embodiments, the indication of whether the realignment procedure has been taken by the UE 22 may be included in Part 1 of the CSI report.

[0138] In some embodiments, the realignment indication from the UE 22 may be a bitfield with a size of 1 bit. I.e., the first value of the bit may represent an indication that the realignment procedure has been taken by the UE 22 and the second value of the bit may represent an indication that the realignment procedure has not been taken by the UE 22. In some embodiments, the bitfield may contain information on the time steps used to produce the CSI report. In other words, the network node 16 may assume that realignment has been taken by the UE 22 when the CSI report transmitted by the UE 22 contains information of time step = 0. For example, consider a maximum time step of 8 that is represented by a 3 -bit indication in Part 1 of the CSI report. Here, 000 may represent time step = 0, 001 may represent time step = 1, 010 may represent time step = 2, and so on. The network node 16 may then realign the historical CSI information according to the network node’s historical CSI information at the respected time step.

[0139] Note that having a bad channel in the previous time step as a reason to conduct the realignment procedure (and indicate to the network node 16 that such action has been taken by the UE 22) may only serve as an example of the possible reasons for the realignment procedure initiated by the UE 22. Other reasons may exist and may also be left to UE 22 proprietary implementation.

[0140] Behavior when the realignment procedure is triggered.

[0141] The realignment procedure may have different mechanisms.

[0142] In some embodiments, the realignment procedure may be in terms of resetting the historical CSI information. In some embodiments, resetting may refer to using partly the same input as the one used to generate the CSI report at time step 0. The partly same input may be the inputs other than the measured channels and its preprocessed derivatives (e.g., eigenvector, W2-like matrix, or non-zero combining coefficients of selected spatial domain basis vectors and frequency domain basis vectors, etc.), such as the states / memory of the model, output from the previous CSI generation, etc. In some embodiments, resetting may refer to zeroing the values. Note that the resetting mechanism may also require the resetting of the time step (e.g., to time step 0). In addition, the UE 22 may also be required to reset the configured timer or counter, if needed.

[0143] In some embodiments, the realignment procedure may be in terms of using the statistics of the historical CSI information. For example, in case the CSI report is missing in the network node 16, and the network node 16 indicates the UE 22 to conduct the realignment procedure, the UE 22 may use the historical CSI information that was used to generate the last CSI information that was successfully received by the network node 16. In another example, the UE 22 (and the network node 16) may use an extrapolation or interpolation mechanism to replace the “unavailable” historical CSI information.

[0144] Some non-limiting example embodiments may include one or more of the following: Embodiment 1. The methods, implemented in a user equipment (UE) capable of implementing Al-based CSI compression with historical CSI, the method comprising: o receiving configuration related to Al-based CSI compression with historical CSI. o receiving either implicit or explicit indication to execute the realignment procedure, o execute the realignment procedure. o [optional] indicate to the network node that a realignment procedure has been taken at the UE side.

[0145] Embodiment 2. Embodiment 1 + where the configuration includes a timer.

[0146] Embodiment 3. Embodiment 1 + where the configuration includes a counter.

[0147] Embodiment 4. Embodiment 1 + where the indication to execute the realignment procedure is the expiration of a predetermined / preconfigured timer.

[0148] Embodiment 5. Embodiment 1 + where the UE executes the realignment procedure when the maximum time step has been achieved.

[0149] Embodiment 6. Embodiment 1 + where the indication of the realignment procedure is at least one of the expiration of the drx-inactivity timer and the start or the expiration of the drx on-duration timer.

[0150] Embodiment 7. Embodiment 1 + where the UE executes the realignment procedure when it receives a DL signal indicating the UE to transmit CSI reports with semi- persistent CSI reporting.

[0151] Embodiment 8. Embodiment 1 + where the UE executes the realignment procedure when it receives a DL signal indicating the UE to transmit CSI reports with aperiodic CSI reporting.

[0152] Embodiment 9. Embodiment 1 + where the UE executes the realignment procedure before it transmits a CSI report with a higher rank compared to the previous CSI report.

[0153] Embodiment 10. Embodiment 1 + where the UE indicates to the network node that the realignment procedure has been taken.

[0154] Embodiment 11. Embodiment 10 + where the indication is included in the CSI report.

[0155] Embodiment 12. Embodiment 11 + where the indication is included in Part 1 of the CSI report.

[0156] Embodiment 13. Embodiment 1 + where the realignment procedure includes at least one of resetting historical CSI information, internal states, and internal memories of the UE-part model. Embodiment 14. Embodiment 13 + where resetting is returning the condition of the model to the condition at time step = 0 or the initial values.

[0157] Embodiment 15. Embodiment 13 + where resetting is zeroing the values of the historical CSI information, the internal states, and the internal memory.

[0158] Embodiment 16. Embodiment 1 + where the realignment is achieved by having the average of the historical CSI information from the earlier time steps.

[0159] Embodiment 17. Embodiment 1 + where the realignment is achieved by extrapolating the historical CSI information from the earlier time steps.

[0160] Embodiment 18. Embodiment 1 + where the indication from the UE includes a bitfield whether the realignment procedure has been taken by the UE.

[0161] Embodiment 19. Embodiment 1 + where the indication from the UE contains information on the time step of the historical CSI information used to produce the CSI report.

[0162] Some embodiments may include one or more of the following examples:

[0163] Example Al. A network node configured to communicate with a user equipment (UE), the network node configured to, and / or comprising a radio interface and / or comprising processing circuitry configured to: transmit a configuration related to artificial intelligence (Al)-based channel state information (CSI) compression with historical CSI; and transmit an indication to execute a realignment procedure based at least in part on the configuration.

[0164] Example A2. The network node of Example Al, wherein the network node, radio interface and / or processing circuitry are further configured to receive an indication that the realignment procedure has been executed.

[0165] Example A3. The network node of any of Examples Al and A2, wherein the configuration includes one of a timer and a counter.

[0166] Example A4. The network node of any of Examples A1-A3, wherein the indication to execute the realignment procedure includes a timer to expire prior to execution of the realignment procedure.

[0167] Example A5. The network node of any of Examples A1-A4, wherein the indication to execute the realignment procedure includes a maximum step time to execution of the realignment procedure. Example A6. The network node of any of Examples A1-A5, wherein the indication to execute the realignment procedure includes a drx-inactivity timer and a drx on-duration timer.

[0168] Example A7. The network node of any of Examples A1-A6, wherein the network node, radio interface and / or processing circuitry are configured to transmit a downlink signal triggering transmission of CSI reports semi-persistently or aperiodically.

[0169] Example A8. The network node of any of Examples A1-A7, wherein the indication to execute the realignment procedure causes the UE to transmit a CSI report with higher rank that a previously reported CSI.

[0170] Example Bl. A method implemented in a network node that is configured to communicate with a user equipment, the method comprising: transmitting a configuration related to artificial intelligence (Al)-based channel state information (CSI) compression with historical CSI; and transmitting an indication to execute a realignment procedure based at least in part on the configuration.

[0171] Example B2. The method of Example Bl, further comprising receiving an indication that the realignment procedure has been executed.

[0172] Example B3. The method of any of Examples Bl and B2, wherein the configuration includes one of a timer and a counter.

[0173] Example B4. The method of any of Examples B1-B3, wherein the indication to execute the realignment procedure includes a timer to expire prior to execution of the realignment procedure.

[0174] Example B5. The method of any of Examples B1-B4, wherein the indication to execute the realignment procedure includes a maximum step time to execution of the realignment procedure.

[0175] Example B6. The method of any of Examples B1-B5, wherein the indication to execute the realignment procedure includes a drx-inactivity timer and a drx on-duration timer.

[0176] Example B7. The method of any of Examples B1-B6, further comprising transmitting a downlink signal triggering transmission of CSI reports semi-persistently or aperiodically.

[0177] Example B8. The method of any of Examples B1-B7, wherein the indication to execute the realignment procedure causes the UE to transmit a CSI report with higher rank that a previously reported CSI. Example Cl. A user equipment (UE) configured to communicate with a network node, the UE configured to, and / or comprising a radio interface and / or processing circuitry configured to: receive a configuration related to artificial intelligence (Al)-based channel state information (CSI) compression with historical CSI; receive an indication to execute a realignment procedure; and execute the realignment procedure based at least in part on the configuration.

[0178] Example C2. The UE of Example Cl, wherein the UE, radio interface and / or processing circuitry are further configured to indicate to the network node that the realignment procedure has been executed.

[0179] Example C3. The UE of any of Examples Cl and C2, wherein the configuration includes one of a timer and a counter.

[0180] Example C4. The UE of any of Examples C1-C3, wherein the indication to execute the realignment procedure includes a timer to expire prior to execution of the realignment procedure.

[0181] Example C5. The UE of any of Examples C 1 -C4, wherein the indication to execute the realignment procedure includes a maximum step time to execution of the realignment procedure.

[0182] Example C6. The UE of any of Examples C 1 -C5, wherein the indication to execute the realignment procedure includes a drx -inactivity timer and a drx on-duration timer.

[0183] Example C7. The UE of any of Examples C1-C6, wherein execution of the realignment procedure occurs upon receiving a downlink signal triggering transmission of CSI reports semi-persistently or aperiodically.

[0184] Example C8. The UE of any of Examples C1-C7, wherein execution of the realignment procedure occurs before transmitting a CSI report with higher rank that a previously reported CSI.

[0185] Example DI. A method implemented in a user equipment (UE) that is configured to communicate with a network node, the method comprising: receiving a configuration related to artificial intelligence (Al)-based channel state information (CSI) compression with historical CSI; receiving an indication to execute a realignment procedure; and executing the realignment procedure based at least in part on the configuration.

[0186] Example D2. The method of Example DI, further comprising indicating to the network node that the realignment procedure has been executed. Example D3. The method of any of Examples DI and D2, wherein the configuration includes one of a timer and a counter.

[0187] Example D4. The method of any of Examples D1-D3, wherein the indication to execute the realignment procedure includes a timer to expire prior to execution of the realignment procedure.

[0188] Example D5. The method of any of Examples D1-D4, wherein the indication to execute the realignment procedure includes a maximum step time to execution of the realignment procedure.

[0189] Example D6. The method of any of Examples D1-D5, wherein the indication to execute the realignment procedure includes a drx-inactivity timer and a drx on-duration timer.

[0190] Example D7. The method of any of Examples D1-D6, wherein execution of the realignment procedure occurs upon receiving a downlink signal triggering transmission of CSI reports semi-persistently or aperiodically.

[0191] Example D8. The method of any of Examples D1-D7, wherein execution of the realignment procedure occurs before transmitting a CSI report with higher rank that a previously reported CSI.

[0192] As will be appreciated by one of skill in the art, the concepts described herein may be embodied as a method, data processing system, computer program product and / or computer storage media storing an executable computer program. Accordingly, the concepts described herein may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects all generally referred to herein as a “circuit” or “module.” Any process, step, action and / or functionality described herein may be performed by, and / or associated to, a corresponding module, which may be implemented in software and / or firmware and / or hardware. Furthermore, the disclosure may take the form of a computer program product on a tangible computer usable storage medium having computer program code embodied in the medium that may be executed by a computer. Any suitable tangible computer readable medium may be utilized including hard disks, CD-ROMs, electronic storage devices, optical storage devices, or magnetic storage devices.

[0193] Some embodiments are described herein with reference to flowchart illustrations and / or block diagrams of methods, systems and computer program products. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, may be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer (to thereby create a special purpose computer), special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0194] These computer program instructions may also be stored in a computer readable memory or storage medium that may direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instruction means which implement the function / act specified in the flowchart and / or block diagram block or blocks.

[0195] The computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0196] It is to be understood that the functions / acts noted in the blocks may occur out of the order noted in the operational illustrations. For example, two blocks shown in succession may in fact be executed substantially concurrently or the blocks may sometimes be executed in the reverse order, depending upon the functionality / acts involved. Although some of the diagrams include arrows on communication paths to show a primary direction of communication, it is to be understood that communication may occur in the opposite direction to the depicted arrows.

[0197] Computer program code for carrying out operations of the concepts described herein may be written in an object oriented programming language such as Python, Java® or C++. However, the computer program code for carrying out operations of the disclosure may also be written in conventional procedural programming languages, such as the "C" programming language. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer. In the latter scenario, the remote computer may be connected to the user's computer through a local area network (LAN) or a wide area network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet Service Provider).

[0198] Many different embodiments have been disclosed herein, in connection with the above description and the drawings. It will be understood that it would be unduly repetitious and obfuscating to literally describe and illustrate every combination and subcombination of these embodiments. Accordingly, all embodiments may be combined in any way and / or combination, and the present specification, including the drawings, may be construed to constitute a complete written description of all combinations and subcombinations of the embodiments described herein, and of the manner and process of making and using them, and may support claims to any such combination or subcombination.

[0199] It will be appreciated by persons skilled in the art that the embodiments described herein are not limited to what has been particularly shown and described herein above. In addition, unless mention was made above to the contrary, it should be noted that all of the accompanying drawings are not to scale. A variety of modifications and variations are possible in light of the above teachings without departing from the scope of the following claims.

Claims

CLAIMS1. A method in a user equipment, UE, (22) configured to communicate with a network node (16), the method comprising: receiving (S26) a configuration related to artificial intelligence, Al, -based channel state information, CSI, compression with historical CSI using a UE part of an Al model; and in response (S28) to a first indication, executing a realignment procedure of the UE part of the Al model based at least in part on the configuration.

2. The method of Claim 1, wherein the configuration includes one of a timer configuration and a counter configuration.

3. The method of any of Claims 1 and 2, wherein the first indication configures the UE (22) to execute the realignment procedure upon expiration of a timer.

4. The method of any of Claims 1-3, wherein the first indication configures the UE (22) to execute the realignment step upon reaching a maximum time step.

5. The method of any of Claims 1-4, wherein the first indication configures the UE (22) to execute the realignment procedure upon at least one of expiration of a drx- inactivity timer, expiration of a drx on-duration timer and start of the drx on-duration timer.

6. The method of any of Claims 1-5, wherein the first indication is included in a downlink signal .

7. The method of Claim 6, wherein the downlink signal is configured to trigger transmission of CSI reports semi-persistently or aperiodically.

8. The method of any of Claims 1-7, wherein execution of the realignment procedure occurs upon receiving a downlink signal triggering transmission of CSI reports semi-persistently or aperiodically.

9. The method of any of Claims 1-8, wherein execution of the realignment procedure occurs before transmitting a CSI report with higher rank than a rank of a previously reported CSI.

10. The method of any of Claims 1-9, wherein the first indication configures the UE (22) to send a second indication to the network node (16) indicating that the realignment procedure has been executed.

11. The method of Claim 10, wherein the second indication that the realignment procedure has been executed is included in the CSI report.

12. The method of any of Claims 10 and 11, wherein the second indication that the realignment procedure has been executed indicates a time step of historical CSI information used to produce a CSI report.

13. The method of any of Claims 1-12, wherein the realignment procedure includes resetting at least one of historical CSI information, internal states and internal memories of a UE part of an Al model.

14. The method of Claim 13, wherein the resetting includes returning a state of the UE part of the Al model to a previous initial state.

15. The method of any of Claims 13 and 14, wherein the resetting includes zeroing values of the historical CSI information, internal states and internal memories of a UE part of an Al model.

16. The method of any of Claims 1-15, wherein execution of the realignment procedure is based at least in part on at least one of averaging and extrapolating from historical CSI information over a plurality of previous time steps.

17. The method of any of Claims 1-16, wherein execution of the realignment procedure is based at least in part on historical CSI information used by the UE (22) to generate a last CSI information successfully received by the network node.

18. A user equipment, UE, (22) configured to communicate with a network node (16), the UE (22) comprising processing circuitry (50) configured to: receive a configuration related to artificial intelligence, Al, -based channel state information, CSI, compression with historical CSI using a UE part of an Al model; and in response to a first indication, execute a realignment procedure of the UE part of the Al model based at least in part on the configuration.

19. The UE (22) of Claim 18, wherein the configuration includes one of a timer configuration and a counter configuration.

20. The UE (22) of any of Claims 18 and 19, wherein the first indication configures the UE (22) to execute the realignment procedure upon expiration of a timer.

21. The UE (22) of any of Claims 18-20, wherein the first indication configures the UE (22) to execute the realignment step upon reaching a maximum time step.

22. The UE (22) of any of Claims 18-21, wherein the first indication configures the UE (22) to execute the realignment procedure upon at least one of expiration of a drx- inactivity timer, expiration of a drx on-duration timer and start of the drx on-duration timer.23 The UE (22) of any of Claims 18-22, wherein the first indication is included in a downlink signal .

24. The UE (22) of Claim 23, wherein the downlink signal is configured to trigger transmission of CSI reports semi-persistently or aperiodically.

25. The UE (22) of any of Claims 18-24, wherein execution of the realignment procedure occurs upon receiving a downlink signal triggering transmission of CSI reports semi-persistently or aperiodically.

26. The UE (22) of any of Claims 18-25, wherein execution of the realignment procedure occurs before transmitting a CSI report with higher rank than a rank of a previously reported CSI.

27. The UE (22) of any of Claims 18-26, wherein the first indication configures the UE (22) to send a second indication to the network node (16) indicating that the realignment procedure has been executed.

28. The UE (22) of Claim 27, wherein the second indication that the realignment procedure has been executed is included in the CSI report.

29. The UE (22) of any of Claims 27 and 28, wherein the second indication that the realignment procedure has been executed indicates a time step of historical CSI information used to produce a CSI report.

30. The UE (22) of any of Claims 18-29, wherein the realignment procedure includes resetting at least one of historical CSI information, internal states and internal memories of a UE part of an Al model.

31. The UE (22) of Claim 30, wherein the resetting includes returning a state of the UE part of the Al model to a previous initial state.

32. The UE (22) of any of Claims 30 and 31, wherein the resetting includes zeroing values of the historical CSI information, internal states and internal memories of a UE part of an Al model.

33. The UE (22) of any of Claims 18-32, wherein execution of the realignment procedure is based at least in part on at least one of averaging and extrapolating from historical CSI information over a plurality of previous time steps.

34. The UE (22) of any of Claims 18-33, wherein execution of the realignment procedure is based at least in part on historical CSI information used by the UE (22) to generate a last CSI information successfully received by the network node.

35. A method implemented in a network node (16) that is configured to communicate with a user equipment, UE, (22), the method comprising: transmitting (S120) a configuration related to artificial intelligence, Al, -based channel state information, CSI, compression with historical CSI;receiving (S22) a CSI report from the UE (22); and executing (S24) a realignment procedure of a network node part of an Al model, based at least in part on the CSI report.

36. The method of Claim 35, wherein the CSI report is based at least in part on a corresponding realignment procedure of a UE part of the Al model.

37. The method of any of Claims 35 and 36, wherein the configuration configures the UE (22) to execute the corresponding realignment procedure upon expiration of a timer.

38. The method of any of Claims 35-37, wherein the CSI report includes an indication that the corresponding realignment procedure has been executed.

39. The method of any of Claims 35-38, wherein the configuration configures the UE (22) to execute the corresponding realignment procedure upon reaching a maximum time step.

40. The method of any of Claims 35-39, wherein the configuration configures the UE (22) to execute the corresponding realignment procedure upon at least one of expiration of a drx-inactivity timer, expiration of a drx on-duration timer and start of the drx on-duration timer.

41. The method of any of Claims 35-40, wherein the CSI report includes a time step of historical CSI information used to produce the CSI report.

42. A network node (16) configured to communicate with a user equipment, UE, (22), the network node (16) comprising processing circuitry (36) configured to: transmit a configuration related to artificial intelligence, Al, -based channel state information, CSI, compression with historical CSI; receive a CSI report from the UE (22); and execute a realignment procedure of a network node part of an Al model, based at least in part on the CSI report.

43. The network node (16) of Claim 42, wherein the CSI report is based at least in part on a corresponding realignment procedure of a UE part of the Al model.

44. The network node (16) of any of Claims 42 and 43, wherein the configuration configures the UE (22) to execute the corresponding realignment procedure upon expiration of a timer.

45. The network node (16) of any of Claims 42-44, wherein the CSI report includes an indication that the corresponding realignment procedure has been executed.

46. The network node (16) of any of Claims 42-45, wherein the configuration configures the UE (22) to execute the corresponding realignment procedure upon reaching a maximum time step.

47. The network node (16) of any of Claims 42-46, wherein the configuration configures the UE (22) to execute the corresponding realignment procedure upon at least one of expiration of a drx -inactivity timer, expiration of a drx on-duration timer and start of the drx on-duration timer.

48. The network node (16) of any of Claims 42-47, wherein the CSI report includes a time step of historical CSI information used to produce the CSI report.

Images