Transmitter node and method for the transmitter node configured to be used in a wireless communication network

Abstract

A Transmitter node (208) that is configured to be used in a wireless communication network (202) including a Network node (204), a Receiver node (206) and the transmitter node is provided. The wireless communication network utilizes Artificial Intelligence, AI, to perform air interface algorithms. The wireless communication network is further configured to obtain channel state information, CSI, and environmental state information, ESI, at a given global time (t), and tokenize them into a state information token (x(t)). It then predicts future token sequences and generates air interface representation inputs by applying representation learning to the channel and environmental properties. The transmitter node predicts future tokens by resampling and interpolating past tokens into an equally spaced sample sequence of state information tokens and uses this information to generate predictions. A FIFO buffer (214, 410, 506) is employed to store and update these samples. The AI algorithms utilize the predicted air interface inputs to optimize network performance.

Description

[0001] TRANSMITTER NODE AND METHOD FOR THE TRANSMITTER NODE CONFIGURED TO BE USED IN A WIRELESS COMMUNICATION NETWORK TECHNICAL FIELD

[0002] The disclosure relates generally to the prediction of Wireless Environment State Information, WESI, tokens and more particularly, the disclosure relates to a Transmitter Node, TX, configured to be used in a wireless communication network to predict a future sequence of state information tokens. Moreover, the disclosure relates to a method for the transmitter node configured to be used in the wireless communication network to predict the future sequence of state information tokens. BACKGROUND

[0003] The wireless environment where communication takes place is primarily sensed through the acquisition of Channel State Information, CSI, which only provides a measure of wireless propagation conditions. Wireless Environment State Information, WESI, represents time-stamped multimodal data describing the wireless propagation environment as illustrated in FIG. 1A. The WESI at time t, denoted as WESI(t), is obtained at the transmitter node, TX, by concatenating channel state information CSI (t) and environment state information ESI(t), at time t, both recorded using a global time reference. WESI(t) is then mapped to a low-dimensional d-dimensional space, with each point referred to as a WESI token.

[0004] The key challenge arises from the fact that air-interface algorithms must make decisions at global time t corresponding to internal sampling index ( ), for establishing communication at a future global time ^corresponding internal sampling index T + P, with only tokens up to global time t being available. The problem that arises is predicting future WESI tokens accurately to improve communication performance.

[0005] Existing approaches in prior art, as illustrated in FIG. 1B, focus on the direct prediction of Channel State Information, CSI, based on previous CSI observations. In FIG. 1B, given a past sequence of M CSI vectors acquired until the internal sampling index T with some time granularity in the internal sampling domain, the future L CSI vectors starting at internal sampling index T + P are predicted. An input to a CSI prediction block 102 is

[0006]

[0007] [ ], i.e., the sequence of M CSI vectors:

[0008]

[0009] [ℓ] = {h[ℓ — M + 1],..., h[ℓ — 1], h[ℓ]}. An output from the CSI prediction block 102 is ĤL[ℓ + P], i.e., the sequence of L CSI vectors containing the predicted CSI for all L resource elements, RE, 104 in a transmission frame starting with sample P + P: HL[P + P] = {h[P + P], h[P + P + 1],..., h[P + P + L — 1]}. The resource element, RE, 104 is a communication resource considered in the standard specification of specified duration TREseconds and is specified frequency bandwidth PREHz. This is the highest / elementary considered granularity (e.g., in time and frequency) in the 3GPP standard. The primary limitation of the prior approaches is the inherent difficulty in predicting CSI, especially in high-dimensional channels that are subject to various impairments.

[0010] To overcome these challenges, there is a need to address the aforementioned technical problems / drawbacks in the prediction of Wireless Environment State Information, WESI, tokens.

[0011] SUMMARY

[0012] It is an object of the disclosure to provide a Transmitter Node, TX, configured to be used in a wireless communication network to predict a future sequence of state information tokens. Moreover, the disclosure relates to a method for the transmitter node configured to be used in the wireless communication network to predict the future sequence of state information tokens. This object is achieved by the features of the independent claims. Further, implementation forms are apparent from the dependent claims, the description, and the figures. The disclosure provides a Transmitter Node, TX, configured to be used in a wireless communication network to predict a future sequence of state information tokens. Moreover, the disclosure relates to a method for the transmiter node configured to be used in the wireless communication network to predict the future sequence of state information tokens.

[0013] According to a first aspect, there is provided a Transmitter Node, TX, configured to be used in a wireless communication network including a Network Node, NW, a Receiver Node, RX, and the Transmitter Node, TX. The wireless communication network is configured to utilize Artificial Intelligence, Al, to perform air interface algorithms. The wireless communication network is further configured to (i) obtain channel state information, CSI, at a global time (t) corresponding to an internal sampling index ( ), wherein the channel state information includes information on channel propagation properties; (ii) obtain environment state information, ESI, at the global time (t) corresponding to an internal sampling index ( ), wherein the environment state information includes information on environmental properties; (iii) tokenize the CSI and ESI together into a state information token (x(t)); (iv) predict a future sequence of state information tokens (y[P + P]) for a future internal sampling index (T + P) based on the past sequence of state information tokens until the global time t (X(t) = {x(t)}); and (v) generate a set of air interface representation inputs (Z\f + P]) for the future internal sampling index by performing representation learning on at least some of the channel properties and environmental properties and the future sequence of state information tokens. Each property on which representation learning is performed on provides one air interface representation input (property i ->

[0014]

[0015] [T + P]). The wireless communication network is then configured to perform the air interface algorithm(s) utilizing Artificial Intelligence air interface algorithm(s) based on the air interface representation inputs (Z[T + P]). The Transmitter Node, TX, is configured to predict the future sequence of state information tokens (y[ + P]) for a future internal sampling index (T + P) based on the past sequence of state information tokens until the global time t (X(t)) by resampling and interpolating the past sequence of state information tokens up to the global time t (X(t) > to an equally spaced sample sequence of state information tokens (yM[P ]) of length AT, wherein the equally-spaced sample past sequence of state information tokens (yM[T] ) comprises the M last samples in the internal sampling domain (yM[ ] = {y[T — M + l],y[T — M + 2],...,y[T]}). Then for each time step (fc) up to the end of future internal sampling index (T + P + L — 1) (fc = 0,..., P + L — 2), the Transmitter Node (TX) is configured to predict a future token (y[T + k + 1]) based on the last M — k samples in the equally-spaced sample past sequence of state information tokens yM[P] ({y[P — M + k + 1],..., y [T]}) and already predicted future tokens {y [T + 1],..., y [T + fc]}. The Transmitter Node (TX) is further configured to store the last M — k samples in the equally-spaced sample past sequence of state information tokens yM[P] ({y[P — M + k + 1],...,y[T]}) and already predicted future tokens {y[T + 1], —,y[-f + fc]} in a First-In-First-Out, FIFO, buffer, and update the buffer as each future token is predicted. After P + L — 1 time steps, the future sequence of state information tokens (7 [T + P] = {y[T + P],y[-t + P + 1],...,y[£ + P + L — 1]}) for the future internal sampling index (T + P) is generated. The transmiter node significantly enhances communication efficiency within the transmission frames by leveraging an AI-based architecture. This architecture processes the Wireless Environment State Information, WESI, tokens generated by the WESI Tokenizer block, which is implemented within the transmiter node. The Al-driven architecture predicts future sequences of WESI tokens, addressing the complex challenge of accurately forecasting Wireless Environment State Information. The predicted tokens are then used as input for a structured WESI representation, facilitating optimal performance of air interface algorithms. The transmiter node employs advanced Al models to predict the WESI sequence through a process that includes both resampling and interpolation, followed by autoregressive prediction. By utilizing historical WESI data and already predicted tokens, the transmiter node generates a future sequence of state information tokens for a specific prediction time. This predictive capability not only enhances the accuracy of environmental state information in wireless networks but also optimizes real-time decision-making for air interface operations.

[0016] Preferably, the transmiter node is further configured to (i) obtain the channel state information, CSI and the environment state information, ESI, at the global time (t) corresponding to the internal sampling index ( ), by obtaining CSI and ESI up to the global time (t), and (ii) tokenize the CSI and ESI together into a past sequence of state information state information tokens up to the global time (t) (JC(t)), and to crop the past sequence of state information tokens up to the global time (t) (JC(t)) to a cropped sequence of state information tokens (JCK(t) ) of K elements up to the global time (t ) i.e., JCK(t) = {x(tK-1'),..., x(t)}. to be used as the sequence of state information tokens up to the global time (t) (JC(t)).

[0017] Optionally, the equally spaced sample sequence of state information tokens yM[£] is at a Resource Element granularity in an internal sampling domain and of length M which is the predictor memory length.

[0018] According to a second aspect, there is provided a method for a transmitter node, TX, configured to be used in a wireless communication network comprising a Network Node, NW, a Receiver Node, RX, and the Transmitter Node, TX. The method includes utilizing Artificial Intelligence, Al, to perform air interface algorithms in the wireless communication network. The method further includes the transmitter node obtaining channel state information, CSI, at a global time (t) corresponding to internal sampling index ( ), wherein the channel state information includes information on channel propagation properties. The method includes obtaining environment state information, ESI, at the global time (t) corresponding to the internal sampling index ( ). The environment state information includes information on environmental properties. The method includes tokenizing the CSI and ESI together into a state information token (x(t)). The method includes predicting a future sequence of state information tokens (y[P + P]) for a future internal sampling index (T + P) based on the past sequence of state information tokens until the global time t (JC(t)). The method includes generating a set of air interface representation inputs (Z[£ + P]) for the future internal sampling index by performing representation learning on at least some of the channel properties and environmental properties and the predicted future sequence of state information tokens. Each property on which representation learning is performed on provides one air interface representation input (property i -> Z^ [T + P] ). The method then includes performing the air interface algorithm(s) utilizing Artificial Intelligence air interface algorithm(s) based on the air interface representation inputs (Z[£ + P]). The method further includes the Transmitter Node, TX, predicting the future sequence of state information tokens (y [T + P] ) for a future internal sampling index (T + P) based on the past sequence of state information tokens until global time t (JC(t)) by (a) resampling and interpolating the cropped sequence of state information tokens (JCK(t)) up to the global time (t) (JC(t)) to an equally spaced sample sequence of state information tokens (yM[^]) °f length M, wherein the sequence of state information tokens (JCK(t) comprises K state information tokens, i.e.,K(t) = {x tx-i,..., x(t) } in a global time reference domain and wherein equally-spaced sample past sequence of state information tokens

[0019]

[0020] ) comprises the AT last samples in an internal sampling domain at the RE granularity (yM[P] = {y[£ — M + l],y[T — M + 2],...,y[T]}), wherein M denotes the predictor memory, and (b) then, for each time step (fc) up to the end of future internal sampling index (T + P + L — 1) (fc = 0,..., P + L — 2), the Transmitter Node (TX) is configured to predict a future token (y[T + fc + 1]) based on the last M — fc samples in the equally-spaced sample past sequence of state information tokens yM[T] ({y [T — M + fc + 1],..., y [T]}) and already predicted future tokens y [T + 1],..., y [T + fc]}, and wherein the Transmitter Node (TX) is further configured to store the last M — fc samples in the equally-spaced sample past sequence of state information tokens yM[P] ( {y[T — M + fc + 1],...,y[T]} ) and already predicted future tokens {y[T + 1],...,y[T + fc]} in a First-In-First-Out, FIFO, buffer, and update the buffer as each future token is predicted. After P + L — 1 time steps, the predicted token sequence y[-f + P] = {y[T + P],y[P + P + 1], —,y[-f + P + L — 1]}) for the future internal sampling index (T + P) is generated.

[0021] The method significantly enhances communication efficiency within the transmission frames by leveraging an Al-based architecture. This architecture processes the Wireless Environment State Information (WESI) tokens generated by the WESI Tokenizer block, which is implemented within the transmitter node. The Al-driven architecture predicts future sequences of WESI tokens, addressing the complex challenge of accurately forecasting Wireless Environment State Information. The predicted tokens are then used as input for a structured WESI representation, facilitating optimal performance of air interface algorithms. The method employs advanced Al models to predict the WESI sequence through a process that includes both resampling and interpolation, followed by autoregressive prediction. By utilizing historical WESI data and already predicted tokens, the method generates a future sequence of state information tokens for a specific prediction time. This predictive capability not only enhances the accuracy of environmental state information in wireless networks but also optimizes real-time decision-making for air interface operations. Preferably, there is provided a computer program product including program instructions for performing the above described method, when the computer program product is executed by one or more processors in a transmitter node.

[0022] Therefore, in contradistinction to the existing solutions, the transmitter node significantly enhances communication efficiency within the transmission frames by leveraging an Al-based architecture. The Al-driven architecture predicts future sequences of WESI tokens, addressing the complex challenge of accurately forecasting Wireless Environment State Information. The transmitter node employs advanced Al models to predict the WESI sequence through a process that includes both resampling and interpolation, followed by autoregressive prediction. By utilizing historical WESI data and already predicted tokens, the transmitter node generates a future sequence of state information tokens for a specific prediction time. This predictive capability not only enhances the accuracy of environmental state information in wireless networks but also optimizes realtime decision-making for air interface operations.

[0023] These and other aspects of the disclosure will be apparent from the implementation(s) described below.

[0024] BRIEF DESCRIPTION OF DRAWINGS

[0025] Implementations of the disclosure will now be described, by way of example only, with reference to the accompanying drawings, in which:

[0026] FIG. 1 A (PRIOR ART) illustrates a typical wireless propagation environment in a wireless communication network;

[0027] FIG. 1B (PRIOR ART) illustrates a typical direct prediction of Channel State Information, CSI, based on previous CSI observations;

[0028] FIG. 2A illustrates a block diagram of a wireless communication network in accordance with an implementation of the disclosure;

[0029] FIG. 2B illustrates the Transmitter Node, TX, of FIG. 2A configured to be used in the wireless communication network in accordance with an implementation of the disclosure;

[0030] FIG. 3 A illustrates a Wireless Environmental State Information, WESI, architecture, implementing acquisition, tokenization, prediction and representation in a wireless communication network in accordance with an implementation of the disclosure; FIG. 3B illustrates a Wireless Environmental State Information, WESI, architecture in accordance with an implementation of the disclosure;

[0031] FIG. 4 illustrates a Wireless Environmental State Information, WESI, sequence predictor block implemented at a Transmitter Node, TX, configured to be used in a wireless communication network in accordance with an implementation of the disclosure; FIG. 5 illustrates a process flow to predict a future sequence of state information tokens (y[ + P]) for a future internal sampling index (T + P) based on the sequence of state information tokens until the global time t (JC(t)) in a wireless communication network in accordance with an implementation of the disclosure; FIG. 6 illustrates a multilayer perception of a WESI token predictor block of a WESI sequence predictor block of in accordance with an implementation of the disclosure;

[0032] FIG. 7 illustrates a Long-Short Term, LSTM, Memory Network predictor of a WESI token predictor block of a WESI sequence predictor block in accordance with an implementation of the disclosure;

[0033] FIG. 8 illustrates a Wireless Environmental State Information, WESI, sequence predictor block implemented in a timefrequency grid configured to be used in a wireless communication network in accordance with an implementation of the disclosure;

[0034] FIG. 9 illustrates an exemplary set of Artificial intelligence, Al, based air interface algorithms implemented in an air interface apparatus in accordance with an implementation of the disclosure;

[0035] FIG. 10 illustrates an architecture of a structured Wireless Environment State Information, WESI, representation in a wireless communication network in accordance with an implementation of the disclosure;

[0036] FIG. 11 illustrates how a Wireless Environment State Information, WESI, structured representation / air interface representation input features are mapped into the time-frequency grid associated with a transmission block in accordance with an implementation of the disclosure;

[0037] FIG. 12 illustrates how ithfeature of a Wireless Environment State Information, WESI, structured representation / air interface representation input mapped into the transmission time-frequency grid in accordance with an implementation of the disclosure; FIG. 13 illustrates an exemplary view of Air interface algorithms’ decisions granularities in accordance with an implementation of the disclosure;

[0038] FIGS. 14A and 14B are flow diagrams that illustrate a method for a transmitter node configured to be used in a wireless communication network in accordance with an implementation of the disclosure; and

[0039] FIG. 15 is an illustration of a computer system in which the various architectures and functionalities of the various previous implementations may be implemented.

[0040] DETAILED DESCRIPTION OF THE DRAWINGS

[0041] Implementations of the disclosure provide a Transmitter Node, TX, configured to be used in a wireless communication network. Moreover, the disclosure relates to a method for the transmitter node configured to be used in the wireless communication network.

[0042] To make solutions of the disclosure more comprehensible for a person skilled in the art, the following implementations of the disclosure are described with reference to the accompanying drawings.

[0043] Terms such as "a first", "a second", "a third", and "a fourth" (if any) in the summary, claims, and foregoing accompanying drawings of the disclosure are used to distinguish between similar objects and are not necessarily used to describe a specific sequence or order. It should be understood that the terms so used are interchangeable under appropriate circumstances, so that the implementations of the disclosure described herein are, for example, capable of being implemented in sequences other than the sequences illustrated or described herein. Furthermore, the terms "include" and "have" and any variations thereof, are intended to cover a non-exclusive inclusion. For example, a process, a method, a system, a product, or a device that includes a series of steps or units, is not necessarily limited to expressly listed steps or units but may include other steps or units that are not expressly listed or that are inherent to such process, method, product, or device.

[0044] Definitions:

[0045] Air Interface: An Air Interface is defined as the specification of technologies and protocols (usually involving algorithms from the physical layer (LI) and data link layer (L2) of the OSI model) enabling the wireless communication between a transmitter node and a receiver node in a wireless communication network using a set of radio resources elements in a time-frequency grid, denoted as a transmission block. Examples of Air Interface algorithms are as follows: precoder design, scheduler design, user localization, beamforming, etc.

[0046] Artificial Intelligence, Al, based air interface: An Al-based air interface refers to an air interface, in which the set of protocols or algorithms implementing the specific wireless functionalities is Al-based. That is, they use pre-trained models based on data to take the corresponding actions / decisions.

[0047] Resource Element: A resource element, RE, is a communication resource considered in the standard specification of specified duration ERE seconds and is specified frequency bandwidth BRE HZ. This is the highest / elementary considered granularity (e.g., in time and frequency) in the 3GPP standard. The following disclosure describes the time and frequency instants using this granularity.

[0048] Transmission block is defined as a set of L • F radio resource elements in the time-frequency grid. A transmission block contains F resource frames, each one including L consecutive resource elements in the time domain. A transmission block contains L resource blocks, each one including F consecutive resource elements in the frequency domain.

[0049] Air interface algorithm: An air interface algorithm or wireless functionality is defined as a process or a set of rules taking decisions / actions enabling the transmission between a transmitter node and a receiver node using a Transmission Block. Note that each wireless functionality may choose a Transmission Block of different size (i.e. different L and F) depending on the needed properties. Indeed, in modem communications systems, the actions performed by a wireless functionality usually depend on the propagation channel over which the communication takes place and the dynamics of this channel is dependent on the considered Resource Block.

[0050] Feature representation learning: In machine learning, feature learning, representation learning, or feature representation learning, is a set of techniques that allows a system to automatically discover the representations needed for feature detection or classification from raw data. This replaces manual feature engineering and allows a machine to both learn the features and use them to perform a specific task. The feature learning is motivated by the fact that machine learning tasks such as classification often require input that is mathematically and computationally convenient to process.

[0051] FIG. 2A illustrates a block diagram of a wireless communication network 202 in accordance with an implementation of the disclosure. The wireless communication network 202 includes a Network Node, NW, 204, a Receiver Node, RX, 206 and the Transmitter Node, TX, 208. The network node 204 serves as a control center within the wireless communication network 202.

[0052] The network node 204 manages network resources, coordinates the communication between the transmitter node 208 and the receiver node 206, and ensures seamless data exchange across the network 202. The network node 204 may also collect Channel State Information, CSI, and Environment State Information, ESI, to optimize the communication process and improve network performance. The transmitter node 208 is responsible for generating and transmitting data signals over the air to other nodes within the network 202. The receiver node 206 is responsible for receiving and decoding the signals transmitted by the transmitter node 208. The receiver node 206 plays a crucial role in ensuring reliable communication, even in challenging environments where signal strength may fluctuate. FIG. 2B illustrates the Transmitter Node, TX, 208 of FIG. 2 configured to be used in the wireless communication network 202 in accordance with an implementation of the disclosure. The wireless communication network 202 is configured to utilize Artificial Intelligence, Al, to perform air interface algorithms. The wireless communication network 202 is further configured to obtain channel state information, CSI, at a global time (t), CSI(t) corresponding to an internal sampling index (T). The channel state information includes information on channel propagation properties. The wireless communication network 202 obtains environment state information, ESI, at the global time (t), ESI(t) corresponding to an internal sampling index (T). The environment state information includes information on environmental properties. The transmitter node 208 integrates a Wireless Environment State Information, WESI, tokenizer block 210 to tokenize the CSI and ESI together into a state information token (x(t)). The WESI(t) is derived from sensors and prior data available at the Network Node, NW, 204, the Receiver Node, RX, 206 and the Transmitter Node, TX, 208 of the wireless communication network 202.

[0053] The transmitter node 208 integrates a WESI sequence predictor block 212 to predict a future sequence of state information tokens (y[P + P]) for a future internal sampling index (T + P) based on the past sequence of state information tokens until the global time t (X(tj), by resampling and interpolating the past sequence of state information tokens up to the global time (t) (X(t) > to an equally spaced sample sequence of state information tokens

[0054]

[0055] ) of length! / . and optionally, X(t) to a cropped sequence of state information tokens XK(t) =

[0056]

[0057] x(t)} including K samples / elements in to the global time reference domain. Optionally, the equally spaced sample sequence of state information tokens is at a Resource Element granularity in an internal sampling domain and of length M which is the predictor memory length. Optionally, the equally-spaced sample past sequence of state information tokens (yM[T]) in the internal sampling domain comprises the! / last samples in the internal sampling domain (yM[ J = {y[^ — M + l],y[T — M + 2],...,y[T]}. For each time step (fc) up to the end of future internal sampling index (T + P + L — 1) (fc = 0,..., P + L — 2), the transmitter node 208 predicts the future token ( y[P + k + 1] ) based on the last M — k samples in the equally-spaced sample past sequence of state information tokens yM[P] ({y[P — M + k + 1],..., y [T]}) and already predicted future tokens y [P + 1],..., y [P + fc]}. After P + L — 1 time steps, the future sequence of state information tokens (y[P + P] = {y[P + P],y[P + P + 1],...,y[P + P + L — 1]}) for the future internal sapling index (P + P) is generated. The transmitter node 208 is further configured to store the last M — k samples in the equally-spaced sample past sequence of state information tokens yM[P] ({y[P — M + k + 1],...,y[T]}) and already predicted future tokens {y [P + 1],..., y [P + fc]} in a First-In-First-Out, FIFO, buffer 214, and update the FIFO buffer 214 as each future token is predicted.

[0058] The transmitter node 208 integrates a WESI structured representation block 216 to generate a set of air interface representation inputs (Z[P + P]) for the future internal sapling index (P + P) by performing representation learning on at least some of the channel properties and environmental properties and the predicted future sequence of state information tokens (y[P + P]). Each property on which representation learning is performed on provides the air interface representation input (property i -> [P + P]). The wireless communication network 202 is then configured to perform the air interface algorithm(s) utilizing Artificial Intelligence air interface algorithm(s) based on the air interface representation inputs (Z[P + P]).

[0059] The transmitter node 208 significantly enhances communication efficiency within the transmission frames by leveraging an Al-based architecture. This architecture processes the Wireless Environment State Information (WESI) tokens generated by the WESI tokenizer block 210, which is implemented within the transmitter node 208. The Al-driven architecture predicts future sequences of WESI tokens, addressing the complex challenge of accurately forecasting Wireless Environment State Information. The predicted tokens are then used as input for a structured WESI representation, facilitating optimal performance of air interface algorithms. The transmitter node 208 employs advanced Al models to predict the WESI sequence through a process that includes both resampling and interpolation, followed by autoregressive prediction. By utilizing historical WESI data and already predicted tokens, the transmitter node 208 generates a future sequence of state information tokens for a specific prediction time. This predictive capability not only enhances the accuracy of environmental state information in wireless networks but also optimizes real-time decision-making for air interface operations.

[0060] Preferably, the transmitter node 208 is further configured to (i) obtain the channel state information, CSI and the environment state information, ESI, at the global time (t) corresponding to the internal sampling index ( ), by obtaining CSI and ESI up to the global time (t), and (ii) tokenize the CSI and ESI together into a past sequence of information tokens (x(t)) up to the global time t (X(t)), and to crop the past sequence of state information tokens up to the global time t (X(t) > to a cropped sequence of state information tokens (XK(t)) of K elements up to the global time t to be used as the sequence of state information tokens (JC(t)).

[0061] FIG. 3 A illustrates a Wireless Environmental State Information, WESI, architecture, implementing acquisition, tokenization, prediction and representation in a wireless communication network in accordance with an implementation of the disclosure. The WESI architecture 300 includes a WESI acquisition block including a channel state information, CSI, acquisition block 302, and an environmental state information, ESI, acquisition block 304, a WESI tokenizer block 306, a WESI sequence predictor block 308, a WESI property list 310, a WESI structured representation block 312, an air interface apparatus 314 and an artificial intelligence, Al, based air interface algorithm 316.

[0062] The CSI acquisition block 302 obtains the CSI at a global time (t). The CSI includes information on channel propagation properties. The ESI acquisition block 304 obtains the ESI at the global time (t). The ESI includes information on environmental property. The environmental property may be at least any one of a camera, a lidar, or weather-related sensors. Optionally, the ESI acquisition block 304 obtains data from prior information available at a transmitter node, TX, a receiver node, RX, and a network node, NW. The prior information may be at least any of a geographical map and a site-specific hardware information, at the network node, the receiver node, and the transmitter node. Optionally, the ESI is a time-synchronized ESI. The ESI acquisition block 304 may be collected, cleaned, timestamped, and stored from the environmental property and the prior information.

[0063] CSI (t) is a time-stamped CSI vector of size F • N obtained either at the transmitter node or the receiver node and containing the frequency-domain propagation channels in h(t), connecting N(t) of the N = NTX• NRXantenna pairs for a subset of F(t) out of the F resources contained in a resource block. The remaining (F — F(t)) • (IV — IV(t)) entries are filled with the special symbol 0csi to indicate missing elements / information.

[0064] ESI(t) is a time-stamped environment state information vector with S = SNW+ SRX+ STXstate information variables, which fuses the environment state information acquired at the NW node, at the RX node and at the TX node, denoted by ESINW(t), ESIRX(t), and ESITX(t), respectively. ESINW(t) is obtained from the local NW node state information, SINW(t), which is a time-stamped vector gathering SNW(t) out of SNWpossible local state information variables (i.e., with 0 ≤ SNW(t) ≤ SNW) acquired at the NW node, by applying data-modality specific tokenizers to the SNW(t) state variables and by filling the remaining (SNW— SNW(t)) state variables with special symbol ØNWto indicate missing information. ESIRX(t) is obtained from the local RX node state information, SIRX(t), which is a time-stamped vector gathering SRX(t) out of SRXpossible local state information variables (i.e., with 0 ≤ SRX(t) ≤ SRX) acquired at the RX node, by applying data-modality specific tokenizers to the SRX(t) state variables and by filling the remaining (SRX— SRX(t)) state variables with special symbol ØRXto indicate missing information. ESITX(t) is obtained from the local TX node state information, SITX(t), which is a time-stamped vector gathering STX(t) out of STXpossible local state information variables (i.e., with 0 ≤ STX(t) ≤ STX) acquired at the TX node, by applying data-modality specific tokenizers to the STX(t) state variables and by filling the remaining (STX— STX(t)) state variables with special symbol ØTXto indicate missing information. The WESI tokenizer block 306 includes a combination of pre-processing and embedding techniques that are configured to (i) handle hardware impairments (e.g., phase offset) in the CSI data and address special symbols in the ESI data, (ii) combine the ESI and CSI data, and (iii) provide an embedding block to map WESI into a lower-dimensional space. The WESI tokenizer block 306 may tokenize the CSI and ESI data into a state information token (x(t)) using a pre-trained model. The WESI tokenizer block 306 applies an embedding process to map the WESI data into a lower-dimensional space. Optionally, x(t) is a token vector of size d < F • N, including a token joint representation of ESI and CSI for all N antenna pairs and all F resource elements within the resource block.

[0065] The WESI sequence predictor block 308 predicts a future sequence of state information tokens (y [F + P] ) for a future internal sampling index (T + P) based on the past sequence of state information tokens until the global time t (JC(t)). The WESI sequence predictor block 308 obtains a length L sequence of state information tokens (y[F + P] ) by predicting the state information tokens for the next transmission frame (of length L) starting at internal sampling index F + P.

[0066] The future sequence of state information tokens (y\F + P)] ) is used to obtain the structured WESI representation (Z[F + P] ) for the future internal sampling index (F + P), which gathers the D individual features (Z1>[F + P] for i = 1,..., D) defined in the WESI property list 310 at a specific time and frequency granularity covering the transmission block starting at the internal sampling index F + P. The WESI property list 310 defines each feature characterizing a physical quantity of a wireless environment. The feature may be a signal strength, an interference level, a channel conditions, and other environmental metrics. The future sequence of state information tokens (y[F + P] ) is represented in minimum time granularity for the whole transmission block starting at internal sampling index F + P. Each wireless environment property in the WESI property list specifies a representation dimensionality such as an output dimension, a representation size, and a time-frequency span of property validity.

[0067] The WESI structured representation block 312 generates a set of air interface representation inputs (Z[F + P] ) for the future internal sampling index (F + P) by performing representation learning on at least some of the channel properties and environmental properties and the future sequence of state information tokens (y[F + P ]). Each property on which representation learning is performed on provides one air interface representation input (property i -> Z^ [F + P] ).

[0068] The air interface apparatus 314 takes decisions on the transmission parameters to enable communication starting at internal sampling index F + P between the transmitter node (e.g., a wireless transmitter node) and the receiver node (e.g., a wireless receiver node) within the transmission block using the Al- based air interface algorithm 316. The transmission block is a set of resource elements in a time-frequency grid. Each of the Al-based algorithms adapt its decision to the wireless channel propagation conditions based on a subset of individual features in the WESI property list 310. The Al-based air interface algorithms 316 may include a precoder design, a scheduler design, a user localization, and a beamforming.

[0069] FIG. 3B illustrates a Wireless Environmental State Information, WESI, architecture in accordance with an implementation of the disclosure. The WESI is derived from sensors and prior data available at a Network Node, NW, a Receiver Node, RX, and a Transmitter Node, TX, of a wireless communication network. The wireless communication network is configured to utilize Artificial Intelligence, Al, to perform air interface algorithms. The WESI includes two main components: (i) channel state information, CSI, and (ii) time-synchronized environment state information, ESI. The WESI data is collected, cleaned, timestamped, and stored from sensors, prior information, and measurement signals. The Artificial Intelligence-based airinterface components may be designed, trained, and inferred using the WESI. The overall WESI architecture is divided into various components, including: (i) a WESI acquisition block, (ii) a WESI tokenization block 320, (iii) a WESI sequence predictor block 322, and (iv) a WESI structured representation block 324. The WESI acquisition block obtains channel state information, CSI, at a global time (t), i.e., CSI(t). The channel state information includes information on channel propagation properties. The WESI acquisition block obtains environment state information, ESI, at the global time (t), ESI(t). The WESI tokenization block 320 is implemented at the transmitter node of the wireless communication network. The CSI (t) and the ESI(t) obtained from the WESI acquisition block are sent as inputs to the WESI tokenization block 320 at the transmitter node. The WESI tokenization block 320 tokenizes the CSI and ESI together into a past sequence of state information tokens (x(t)) up to the global time (t)

[0070]

[0071] The x(t) is a token vector of size d < F • N, including a token joint representation of ESI and CSI for all N antenna pairs and all F resource elements within the resource block.

[0072] The WESI sequence predictor block 322 predicts a future sequence of state information tokens (y[ ] ) for a future internal sampling index (T + P) based on the past sequence of state information tokens up to the global time (t) ( JC( t) ). The WESI structured representation block 324 generates a set of air interface representation inputs (Z[£ + P] ) for the future internal sampling index (T + P) by performing representation learning on at least some of the channel properties and environmental properties and the future sequence of state information tokens (y[F + P] ). Each property on which representation learning is performed on provides one air interface representation input (property i -> Z^[f + P]). The wireless communication network is then configured to perform the air interface algorithm(s) utilizing Al air interface algorithm(s) based on the air interface representation inputs (Z[£ + P]).

[0073] FIG. 4 illustrates a Wireless Environmental State Information, WESI, sequence predictor block 402 implemented at a Transmitter Node, TX, configured to be used in a wireless communication network in accordance with an implementation of the disclosure. The wireless communication network is configured to utilize Artificial Intelligence, Al, to perform air interface algorithms. The WESI sequence predictor block 402 is implemented at a transmitter node of the wireless communication network. The WESI is derived from sensors and prior data available at a network node, a receiver node and the transmitter node of the wireless communication network. The WESI includes two main components: (i) channel state information, CSI, and (ii) time-synchronized environment state information, ESI. The transmitter node of the wireless communication network is configured to obtain channel state information, CSI, at a global time (t), CSI(t). The channel state information includes information on channel propagation properties. The transmitter node obtains environment state information, ESI, at the global time (t), ESI(t). The environment state information includes information on environmental properties. The transmitter node integrates a Wireless Environment State Information, WESI, tokenizer block to tokenize the CSI and ESI together into state information tokens (x(t)) up to the global time (t).

[0074] The WESI sequence predictor block 402 includes a WESI token sequencing block 404 and a WESI token predictor block 406.

[0075] The WESI token sequencing block 404 takes the past state information tokens (x(t)) (i.e., WESI tokens) derived by the WESI tokenization block at the transmitter node up to the global time t (JC(t)) as input, and crops the past sequence of state information tokens t (JC(t)) to a cropped sequence of state information tokens (JCK(t)) of K elements up to the global time t. The WESI token predictor block 406 takes an equally-spaced sample past sequence of state information tokens yM[f] in the internal sampling domain obtained from the WESI token sequencing block 404 as input. The WESI token predictor block 406 includes a WESI predictor 408 and a First-In-First-Out, FIFO, buffer 410. The WESI token predictor block 406 predicts the future sequence of state information tokens 7 [T + P] iteratively using P + L — 1 iterations. At iteration 0, the WESI predictor 408 takes as input sequence yM[ ] and predicts token y [T + 1]. Then, at iteration fc, 0 < fc < P + L — 1, the FIFO buffer 410 of capacity M collects the concatenation of the last M — k samples of yM[F] ({y[^ — M + fc + 1],...,y[T]})and the already predicted token sequence {y[T + l],

[0076]

[0077] + fc]}, providing as input to the WESI predictor 408 {y[T — M + k + 1],...,y[T],y[T + 1],...,y[£ + k]}. At iteration k, the WESI token predictor block 406 stores the last M — k samples of 7M[T] {y[^—M + k + 1],...,y[T]} and the already predicted token sequence {y[T + 1], —,y[-f + fc]} in the FIFO buffer 410, and updates the FIFO buffer 410 as each future token is predicted. After (P + L — 1) iterations, the WESI predictor 408 outputs the predicted future sequence of state information tokens y[-f + P] = {y[-f + P],y[T + P + 1], —,y[-f + P + L — 1]}.

[0078] FIG. 5 illustrates a process flow to predict a future sequence of state information tokens (y\P + P]) for a future internal sampling index (P + P) based on the past sequence of state information tokens t (X(t) > in a wireless communication network in accordance with an implementation of the disclosure. The wireless communication network is configured to utilize Artificial Intelligence, Al, to perform air interface algorithms. A transmitter node of the wireless communication network is configured to obtain channel state information, CSI, at a global time (t), CSI(t). The channel state information includes information on channel properties. The transmitter node obtains environment state information, ESI, at the global time (t), ESI(t). The environment state information includes information on environmental properties. The transmitter node integrates a Wireless Environment State Information, WESI, tokenizer block to tokenize the CSI and ESI together into state information tokens (x(t)) up to the global time (t).

[0079] The transmitter node of the wireless communication network is deployed with a WESI sequence predictor block to predict a future sequence of state information tokens (y[P + P]) for a future internal sampling index (T + P) based on the sequence of state information tokens (JC(t)). The WESI sequence predictor block includes a WESI token sequencing block 502 and a WESI token predictor block 504.

[0080] At step 510, at a global time t, the state information token x(t) (i.e., WESI token) that is produced by the WESI tokenization block is provided to the WESI token sequencing block 502. Each state information token (x(t)) is derived from time-stamped multimodal data that reflects the wireless propagation environment, including channel state information, CSI, and environment state information, ESI. At step 512, the sequence of state information tokens including all WESI tokens x(t) is collected up to the global time (t) (JC(t)) to a cropped sequence of state information tokens (JCK(t)) of K elements up to the global time (t), i.e., XK(fi) = {x(tK-t) x(t)}.

[0081] At step 514, the cropped sequence of state information tokens (XK(t)) is resampled and interpolated to obtain a sequence of M equally-spaced tokens in the internal sampling domain. This resampling adapts the global time reference sequence into an internal sampling domain. The result is a sequence of tokens i'l / M|f,|) that are uniformly spaced and ready for prediction. Optionally, the equally-spaced sample sequence of state information tokens is at the Resource Element, RE, granularity in the internal sampling domain and of length M which is the predictor memory length, i.e., (yM[P] = {y[T — M + l],y[T — M + 2],..., y [T]}). At step 516, the M tokens are then passed to the WESI token predictor block 504, which iteratively predicts the next P + L samples. This is achieved through a deep learning model or algorithm, which analyzes the pattern and progression of the tokens. For example, for each time step (fc) up to the end of future internal sampling index (T + P + L — 1) (fc = 0,...,? + L — 2), the WESI token predictor block 504 stores the last M — k samples in the equally-spaced sample past sequence of state information tokens yM[P] ( {y[T — M + k + 1],...,y[P]} ) and already predicted future tokens {y[T + 1],...,y[T + &]} in a First-In-First-Out, FIFO, buffer, 506 and updates the FIFO buffer 506 as each future token is predicted. At step 518, for each time step / iteration, fc, up to the end of future internal sampling index (T + P + L — 1) (fc = 0,..., P + L — 2), the WESI predictor 508 predicts a future token (y[P + k + 1]) based M samples stored in the FIFO buffer 506. At step 520, after P + L — 1 time steps, the future sequence of state information tokens ( y\P + P\ = {y[P + Pj.yfP + P + 1]> ■■■,y[P + P + L — 1]}) for the future internal sampling index (T + P) is produced as the output, characterizing the transmission resource block starting at time T + P given the past WESI tokens until global time t, which may be used for further decision-making in the communication network. This sequence represents the anticipated wireless environmental state, which can be used by the communication system to enhance decision-making, optimize network performance, and adapt transmission strategies. FIG. 6 illustrates a multilayer perception of a WESI token predictor block of a WESI sequence predictor block in accordance with an implementation of the disclosure. The WESI sequence predictor block includes a WESI token sequencing block and a WESI token predictor block 602. The WESI token predictor block 602 may be implemented, for instance, using one of the following specific architectures: (i) a Multilayer Perceptron architecture (as illustrated in FIG. 6), (ii) a Long-Short Term Memory architecture (as illustrated in FIG. 7) or (iii) a Kalman filter architecture. FIG. 7 illustrates a Long-Short Term, LSTM, Memory Network predictor of a WESI token predictor block 702 of a WESI sequence predictor block in accordance with an implementation of the disclosure.

[0082] FIG. 8 illustrates a Wireless Environmental State Information, WESI, sequence predictor block 804 implemented in a timefrequency grid configured to be used in a wireless communication network in accordance with an implementation of the disclosure. The wireless communication network includes a Network Node, NW, a Receiver Node, RX, and the Transmitter Node, TX, a global time reference domain. The WESI token sequencing block 806 applies cropping to and selection 806A to the past sequence of state information tokens up to the global time (t) (X(t) > to obtain the cropped sequence of state information tokens (XK(t)) and resampling and interpolation 806B to the cropped sequence of state information tokens (XK(t) = {x(tK-1), x(t)}) in a global time reference domain 802 to obtain a equally-spaced sample past sequence of state information tokens

[0083]

[0084] = {y[f — M + l],y[T — M + 2],...,y[T]} of M equally spaced samples at a Resource Element, RE, granularity in an internal sampling domain 810, where M denotes is predictor memory (i.e., the length). The WESI token predictor block 808 takes the equally-spaced sample past sequence of state information tokens yM[ ] obtained from the WESI token sequencing block 806 as input in the internal sampling domain 810. The WESI token predictor block 808 includes a WESI predictor and a First-In-First-Out, FIFO, buffer. The WESI token predictor block 808 predicts the future sequence of state information tokens y[-f + P] = {y[T + P],y[ + P + 1], —,y[-f + P + L]} iteratively using P + L — 1 iterations. The future sequence of state information tokens y [T + P] contains tokens from time T + P up to the end of future internal sampling index (T + P + L — 1).

[0085] FIG. 9 illustrates an exemplary Artificial intelligence, Al, based air interface algorithms implemented in an air interface apparatus in accordance with an implementation of the disclosure. The future sequence of state information tokens 902 (y [T + P]) containing predicted WESI tokens is given as input to the WESI Structured Representation Block 904 including D WESI feature representation learning modules, each one corresponding to a one of D properties contained in the WESI property list 908 (property i -> Z^[f + P]). FIG. 9 explicitly shows a WESI feature 1 representation learning module 904A, a WESI feature 2 representation learning module 904B, a WESI feature i representation learning module 9041, a WESI feature D — 1 representation learning module 904D, and a WESI feature D representation learning module 904C.

[0086] A possibly different subset of the set of air interface representation inputs (Z(T + P)) are combined to generate an air interface representation input to each one of the A Al-based air interface algorithms 906 to make decisions on transmission parameters for communication. For instance, in FIG. 9 the output of the WESI feature 1 representation learning module 904 A represented as Z^1' [T + P] and the WESI feature 2 representation learning module 904B represented as Z2>[T + P] combines with the output of the WESI feature D — 1 representation learning module 904D represented as

[0087]

[0088] |f + P] to generate an air interface representation input and communicates the air interface representation input to an Al-based air interface algorithm 906A to make decisions on transmission parameters for communication. The output of the WESI feature i representation learning module 9041 represented as ZLI[T + P] generates and sends an air interface representation input to an Al based air interface algorithm 406C to make decisions on transmission parameters for communication. The output of the WESI feature D representation learning module 904C represented as ZDl[T + P] and WESI feature D — 1 representation learning module 904D represented as

[0089]

[0090] l![T + P] are combined to generate an air interface representation input and communicates the air interface representation input to the Al based air interface algorithm 906B to make decisions on transmission parameters for communication.

[0091] FIG. 10 illustrates an architecture for Structured Wireless Environment State Information, WESI, representation in accordance with an implementation of the disclosure. The architecture shows a predicted future sequence of state information tokens 1002 (1 / [ + P]) containing predicted WESI tokens, a WESI property list 1004, and a WESI structured representation block 1006.

[0092] The WESI property list 1004 includes D WESI Properties (explicitly in the figure, a WESI Property 1, a WESI Property i, and a WESI Property D). Each WESI property specifies a representation dimensionality such as an output dimension, a representation size, and a time-frequency span of property validity associated with the corresponding WESI feature representation learning module.

[0093] The WESI structured representation block 1006 includes a WESI feature 1 representation learning module 1006A, a WESI feature i representation learning module 10061, and a WESI D representation learning module 1006D. The future sequence of state information tokens 1002 (y [T + P] ) is transferred to the D WESI feature representation learning modules for performing representation learning on the property to each one of the WESI properties specified in the WESI property list 1004 and to generate a set of air interface representation inputs. Such air interface representation input is provided as input to an Al-based air interface algorithm. For instance, for the predicted future sequence of state information tokens y[£ + P], containing the L predicted WESI tokens at minimum time granularity for a transmission block starting at an internal sampling index T + P, the WESI structured representation block 1006 obtains the air interface representation input (i.e., WESI structured representation) as Z\-f>+ P] = + P],..., Zlli\-f' + P],..., Z!i>i\-f' + P]}, where Zlli\-f' + P] denotes the ith individual feature, which corresponds to the ith WESI property specified in the WESI property list 1004 for the transmission block starting at time T + P at its specified time and frequency granularity as shown in FIG. 11.

[0094] FIG. 11 illustrates how a Wireless Environment State Information, WESI, structured representation / air interface representation input features are mapped into the time-frequency grid associated with a transmission block in accordance with an implementation of the disclosure. For example, the WESI structured representation of feature 1 1102 includes a time granularity of 4 and a frequency granularity of F / 4. Hence, the WESI representation learning of feature 1 is

[0095]

[0096] {z® [F + P], z® + P |,...,t[T + P], z® + P |,..., z®44[F + P] j. Similarly, the WESI structured representation of feature D 1104 includes a time granularity of L RE and a frequency granularity of F / 2 REs. Hence, the WESI representation learning of feature D 1104 is {z® [ + P], z^ [T + P] j.

[0097] FIG. 12 illustrates how ithfeature of a Wireless Environment State Information, WESI, structured representation / air interface representation are mapped into the time-frequency grid associated with a transmission block in accordance with an implementation of the disclosure. The WESI structured representation of feature i 1202 has time granularity r;= L / Lt, where L is the total number of REs in a resource frame, and frequency granularity 2,- = F / Ftwhere F is total number of REs in a resource block. Then, the WESI structured representation is obtained asL)\

[0098]

[0099] f + P] = [T + Pj.z^t / + P],..., Zp*)1[ / + P], z® [T + P],

[0100]

[0101] , Zp^L. [T + P] J, where z®. [T + P] G is the feature vector characterizing the ith WESI property for time frequency block containing F;x Ltresource elements starting at the resource element at position ( / — 1)F;+ 1 of the resource block at position (j —!)£( + 1 in the time of the transmission block.

[0102] FIG. 13 illustrates an exemplary view of an Air interface algorithm’s decisions granularities in accordance with an implementation of the disclosure. A set of A Al-based air interface algorithms 1302 take decision on the transmission parameters for a transmission block to be used for communication. Each air interface algorithm 1302 has a time granularity and a frequency granularity to take decisions on non-overlapping time frequency blocks inside the transmission block. The time and frequency granularities of a WESI structure representation required by an air interface algorithm 1302 are not smaller than the time and frequency granularities of the algorithm’s decisions.

[0103] FIGS. 14A-14B are flow diagrams that illustrate a method for a transmitter node configured to be used in a wireless communication network in accordance with an implementation of the disclosure. The wireless communication network includes a Network Node, NW, a Receiver Node, RX, and the Transmitter Node, TX. The method utilizes Artificial Intelligence, Al, to perform air interface algorithms in the wireless communication network. At step 1402, the method includes the transmitter node obtaining channel state information, CSI, at a global time (t) corresponding to internal sampling index (T). The channel state information includes information on channel propagation properties. At step 1404, environment state information, ESI, at the global time (t) corresponding to the internal sampling index ( ) is obtained. The environment state information includes information on environmental properties. At step 1406, the CSI and ESI is tokenized together into a state information token (x(t)). At step 1408, a future sequence of state information tokens (y[ + P]) is predicted for a future internal sampling index (T + P) based on the past sequence of state information tokens until the global time t ({x(t)}). At step 1410, a set of air interface representation inputs (Z[£ + P)]) is generated for the future internal sampling index by performing representation learning on at least some of the channel properties and environmental properties and the predicted future sequence of state information tokens. Each property on which representation learning is performed on provides one air interface representation input (property i -> 2®[]). At step 1412, the air interface algorithm(s) is performed by utilizing Artificial Intelligence, Al, air interface algorithm(s) based on the air interface representation inputs (Z[£ + P)]). The method further comprise the Transmitter Node (TX) predicting the future sequence of state information tokens ('Z [T + P]) for a future internal sampling index (T + P) the past sequence of state information tokens until the global time t ({x(t)}) by resampling and interpolating the sequence of state information tokens (JCK(t)) up to the global time (t) (JC(t)) to an equally-spaced sample sequence of state information tokens (yM[£ ] ) of length M, wherein the sequence of state information tokens (JCK(t)) comprises K state information tokens and wherein equally-spaced sample past sequence of state information tokens (Y ) comprises the AT last samples in the internal sampling domain (yMUr] = {y[^—W + 1],...,y[T]}). Then for each time step (fc) up to the end of future internal sampling index (T + P + L — 1) (fc = 0,..., P + L — 2), the method predicts the future token (y[T + k + 1]) based on the last M — i samples in the equally-spaced sample past sequence of state information tokens 7M[T] ({y[T — M + fc + 1],...,y[T]}) and already predicted future tokens {y[T + 1],...,y[-f + fc]}. After P + L — 1 time steps, the future sequence of state information tokens (7 [T + P] = {y[T + P],y[P + P + 1],...,y[P + P + L — 1]}) for the future internal sampling index (T + P) is generated. The method is further configured to store the last M — k samples in the equally-spaced sample past sequence of state information tokens yM[ ] ({y [T — M + k + 1],..., y [T]}) and already predicted future tokens {y[T + 1], ■■■,y[P + fc]} in a First-In-First-Out, FIFO, buffer, and updates the buffer as each future token is predicted.

[0104] The method significantly enhances communication efficiency within the transmission frames by leveraging an Al-based architecture. This architecture processes the Wireless Environment State Information (WESI) tokens generated by the WESI Tokenizer block, which is implemented within the transmitter node. The Al-driven architecture predicts future sequences of WESI tokens, addressing the complex challenge of accurately forecasting Wireless Environment State Information. The predicted tokens are then used as input for a structured WESI representation, facilitating optimal performance of air interface algorithms.

[0105] The method employs advanced Al models to predict the WESI sequence through a process that includes both resampling and interpolation, followed by autoregressive prediction. By utilizing historical WESI data and already predicted tokens, the method generates a future sequence of state information tokens for a specific prediction time. This predictive capability not only enhances the accuracy of environmental state information in wireless networks but also optimizes real-time decision-making for air interface operations. FIG. 15 is an illustration of a computer system in which the various architectures and functionalities of the various previous implementations may be implemented. As shown, the computer system 1500 includes at least one processor 1504 that is connected to a bus 1502, wherein the computer system 1500 may be implemented using any suitable protocol, such as PCI (Peripheral Component Interconnect), PCI-Express, AGP (Accelerated Graphics Port), Hyper Transport, or any other bus or point-to-point communication protocol (s). The computer system 1500 also includes a memory 1506.

[0106] Control logic (software) and data are stored in the memory 1506 which may take a form of random-access memory (RAM). In the disclosure, a single semiconductor platform may refer to a sole unitary semiconductor-based integrated circuit or chip. It should be noted that the term single semiconductor platform may also refer to multi-chip modules with increased connectivity which simulate on-chip modules with increased connectivity which simulate on-chip operation, and make substantial improvements over utilizing a conventional central processing unit (CPU) and bus implementation. Of course, the various modules may also be situated separately or in various combinations of semiconductor platforms per the desires of the user.

[0107] The computer system 1500 may also include a secondary storage 1510. The secondary storage 1510 includes, for example, a hard disk drive and a removable storage drive, representing a floppy disk drive, a magnetic tape drive, a compact disk drive, digital versatile disk (DVD) drive, recording device, universal serial bus (USB) flash memory. The removable storage drive at least one of reads from and writes to a removable storage unit in a well-known manner.

[0108] Computer programs, or computer control logic algorithms, may be stored in at least one of the memory 1506 and the secondary storage 1510. Such computer programs, when executed, enable the computer system 1500 to perform various functions as described in the foregoing. The memory 1506, the secondary storage 1510, and any other storage are possible examples of computer-readable media.

[0109] In an implementation, the architectures and functionalities depicted in the various previous figures may be implemented in the context of the processor 1504, a graphics processor coupled to a communication interface 1512, an integrated circuit (not shown) that is capable of at least a portion of the capabilities of both the processor 1504 and a graphics processor, a chipset (namely, a group of integrated circuits designed to work and sold as a unit for performing related functions, and so forth). Furthermore, the architectures and functionalities depicted in the various previous-described figures may be implemented in a context of a general computer system, a circuit board system, a game console system dedicated for entertainment purposes, an application-specific system. For example, the computer system 1500 may take the form of a desktop computer, a laptop computer, a server, a workstation, a game console, an embedded system.

[0110] Furthermore, the computer system 1500 may take the form of various other devices including, but not limited to a personal digital assistant (PDA) device, a mobile phone device, a smart phone, a television, and so forth. Additionally, although not shown, the computer system 1500 may be coupled to a network (for example, a telecommunications network, a local area network (LAN), a wireless network, a wide area network (WAN) such as the Internet, a peer-to-peer network, a cable network, or the like) for communication purposes through an I / O interface 1508.

[0111] It should be understood that the arrangement of components illustrated in the figures described are exemplary and that other arrangement may be possible. It should also be understood that the various system components (and means) defined by the claims, described below, and illustrated in the various block diagrams represent components in some systems configured according to the subject matter disclosed herein. For example, one or more of these system components (and means) may be realized, in whole or in part, by at least some of the components illustrated in the arrangements illustrated in the described figures. In addition, while at least one of these components are implemented at least partially as an electronic hardware component, and therefore constitutes a machine, the other components may be implemented in software that when included in an execution environment constitutes a machine, hardware, or a combination of software and hardware.

[0112] Although the disclosure and its advantages have been described in detail, it should be understood that various changes, substitutions, and alterations can be made herein without departing from the spirit and scope of the disclosure as defined by the appended claims.

Claims

CLAIMS1. A Transmitter Node, TX, (208) configured to be used in a wireless communication network (202) comprising a Network Node, NW, (204) a Receiver Node, RX, (206) and the Transmitter Node, TX, (208) wherein the wireless communication network is configured to utilize Artificial Intelligence to perform air interface algorithms, wherein the wireless communication network is characterized in that the wireless communication network is further configured to:obtain channel state information, CSI, at a global time (t) corresponding to an internal sampling index ( ), wherein the channel state information comprises information on channel propagation properties,obtain environment state information, ESI, at the global time (t) corresponding to the internal sampling index ( ), wherein the environment state information comprises information on environmental properties,tokenize the CSI and ESI together into a state information token (x(t)),predict a future sequence of state information tokens (y[P + P]) for a future internal sampling index (T + P) based on the past sequence of state information tokens until the global time t ({x(t)}),generate a set of air interface representation inputs (Z[£ + P)]) for the future internal sampling index (T + P) by performing representation learning on at least some of the channel properties and environmental properties and the future sequence of state information tokens, wherein each property on which representation learning is performed on provides one air interface representation input (property i -> Z^ [T + P] ), and then to perform the air interface algorithm(s) utilizing Artificial Intelligence air interface algorithm(s) based on the air interface representation inputs (Z[£ + P]),wherein the Transmitter Node (TX) is configured to predict the future sequence of state information tokens (If + f] ) forafuture internal sampling index (T + P) based on the past sequence of state information tokens until the global time t ({x(t)}) byresampling and interpolating the past sequence of state information tokens up to the global time (t) (JC(t)) to an equally-spaced sample past sequence of state information tokensof length M, wherein the equally-spaced sample past sequence of state information tokenscomprises the M last samples in the internal sampling domain (yM[T] = {y [T — M + 1], y [T — M + 2],..., y [T]}), and then for each time step (fc) up to the end of future internal sampling index (T + P + L — 1) (fc = 0 P +L - 2),predict a future token (y[T + k + 1]) based on the last M — k samples in the equally-spaced sample past sequence of state information tokens yM[T] ({y [T — M + k + 1],..., y [T]}) and already predicted future tokens {y[T + 1],...,y[T + fc]} to generate the future sequence of state information tokens (7 [T + P] = {y[T + P],y[T + P], ■■■,y[-t + P + L — 1]}) for the future internal sampling index (T + P), andwherein the Transmitter Node (TX) is further configured to store the last M — k samples in the equally- spaced sample past sequence of state information tokens yM[P] ({y[T — M + k + 1],...,y[T]}) andalready predicted future tokens {y\f' + 1],...,y[-£ + fc]} in a First-In-First-Out, FIFO, buffer, (214, 410, 506) and update the buffer as each future token is predicted.

2. The transmitter node (208) according to claim 1, wherein the Transmitter Node (TX) is further configured to: obtain the channel state information, CSI and the environment state information, ESI, at the global time (t) corresponding to the internal sampling index ( ), by obtaining CSI and ESI up to the global time (t), and tokenize the CSI and ESI together into the past sequence of state information tokens up to the global time (t) (JC(t)), and tocrop the past sequence of state information tokens up to the global time (t) (JC(t)) to a cropped sequence of state information tokens (XK(t)) of K elements up to the global time (t) to be used as the past sequence of state information tokens of state information tokens up to the global time (t) (JC(t)).

3. The transmitter node (208) according to any preceding claim, wherein the equally spaced sample sequence of state information tokens yM[£ ]) is at a Resource Element granularity in an internal sampling domain and of length M which is the predictor memory length.

4. A method for a transmitter node (208) configured to be used in a wireless communication network (202) comprising a Network Node, NW, (204) a Receiver Node, RX (206) and the Transmitter Node, TX, (208), wherein the method comprises utilizing Artificial Intelligence to perform air interface algorithms in the wireless communication network, wherein the method is characterized in that the method further comprises the transmitter node:obtaining channel state information, CSI, at a global time (t) corresponding to internal sampling index ( ), wherein the channel state information comprises information on channel propagation properties,obtaining environment state information, ESI, at the global time (t) corresponding to the internal sampling index ( ), wherein the environment state information comprises information on environmental properties,tokenizing the CSI and ESI together into a state information token (x(t)),predicting a future sequence of state information tokens (y[-t + P]) for a future internal sampling index (T + P) based on the past sequence of state information tokens until the global time t ({x(t)}),generating a set of air interface representation inputs (Z[£ + P]) for the future internal sampling index (T + P) by performing representation learning on at least some of the channel properties and environmental properties and the predicted future sequence of state information tokens, wherein each property on which representation learning is performed on provides one air interface representation input (property i -> Z^ [T + P] ), and then performing the air interface algorithm(s) utilizing Artificial Intelligence air interface algorithm(s) based on the air interface representation inputs (Z[£ + P]),wherein the method further comprises the Transmitter Node, TX, predicting the future sequence of state information tokens (y[ + P]) for a future internal sampling index (T + P) based on the past sequence of state information tokens until the global time t ({x(t)}) byresampling and interpolating the sequence of state information tokens up to the global time (t) to an equally spaced sample sequence of state information tokens (yM[P ]) of length M, wherein the equally-spaced sample past sequence of state information tokens yM[P] ) comprises the M last samples in the internal sampling domain (yM[P] = {y[P — M + l],y[P — M + 2],...,y[T ]}), and then for each time step (fc) up to the end of future internal sampling index (P + P + L — 1) (fc = 0,..., P + L — 2),predicting a future token (y[P + fc + 1]) based on the last M — fc samples in the equally-spaced sample past sequence of state information tokens yM[P] ({y[ — Af + fc + 1],...,y[ ]}) and already predicted future tokens {y[P + 1],...,y[P + fc]} to generate the future sequence of state information tokens (y[P + P] = {y[P + P],y[P + P + 1],...,y[T + P + L — 1]}) for the future internal sampling index (P + P), and wherein the Transmitter Node (TX) is further configured to store the last M — fc samples in the equally-spaced sample past sequence of state information tokens yM[P] ({y[P — M + fc + 1],...,y[T]}) and already predicted future tokens {y[P + 1],...,y[P + fc]} inaFirst-In-First-Out, FIFO, buffer, (214, 410, 506) and update the buffer as each future token is predicted.

5. A computer program product comprising program instructions for performing the method according to claim 4, when executed by one or more processors in a transmitter node.

Images