Waveform switching without signalling

Abstract

Method comprising selecting a first waveform among a set of plural waveforms to be used for a first transmission from a terminal to a base station based on assistance information; instructing a transmitter of the terminal to perform the first transmission to the base station using the first waveform; wherein the assistance information comprises a channel condition of a channel between the terminal and the base station.

Description

[0001]WAVEFORM SWITCHING WITHOUT SIGNALLING Technical Field The present disclosure relates to waveform switching. Background In 3GPP (3rdgeneration partnership project) 5G (5thgeneration) NR (New Radio), a UL (uplink) transmission may be performed using one of two different OFDM (orthogonal frequency division multiplexing) waveforms: Discrete Fourier Transform spread OFDM (DFT-S-OFDM) (also known as Single-Carrier OFDM (SC-OFDM) or Linearly-Precoded OFDM Access (LP- OFDMA)) and Cyclic-prefix OFDM CP-OFDM. The base station (next generation NodeB, gNB) controls the waveform used by the terminal (user equipment, UE) by a DCI (downlink control information) message. Summary It is an object to improve the prior art. According to a first aspect, there is provided an apparatus comprising means for selecting a first waveform among a set of plural waveforms to be used for a first transmission from a terminal to a base station based on assistance information; means for instructing a transmitter of the terminal to perform the first transmission to the base station using the first waveform; wherein the assistance information comprises a channel condition of a channel between the terminal and the base station. The means for selecting may be configured to select the first waveform without using an instruction of the waveform to be used received from the base station. The apparatus may further comprise means for inhibiting the terminal to inform the base station that the first waveform is selected for the first transmission. (EN_00 Anlage mit Briefkopf.dot-HL211 / MG162 / RM158) The apparatus may further comprise means for receiving an instruction on hysteresis from the base station; means for prohibiting the means for selecting to select a second waveform different from the first waveform for a second transmission following the first transmission if a condition defined by the instruction on the hysteresis is not fulfilled. The apparatus may further comprise means for receiving, from the base station, a prohibition to select the first waveform based on the assistance information and an instruction to use a third waveform for the first transmission; means for prohibiting the means for selecting from selecting the first waveform based on the assistance information in response to receiving the prohibition; wherein the means for instructing may be configured to instruct the transmitter of the terminal to perform the first transmission using the third waveform in response to receiving the instruction to use the third waveform. The assistance information may comprise a history of waveforms previously used for respective one or more transmissions from the terminal to the base station. The apparatus may further comprise means for checking whether the first waveform is different from a waveform used for a transmission from the terminal to the base station preceding the first transmission; means for informing the base station that the first waveform is different from the waveform used for the transmission from the terminal to the base station preceding the first transmission in response to checking that the first waveform is different from the waveform used for the transmission from the terminal to the base station preceding the first transmission. The assistance information may comprise at least one of a downlink signal to interference and noise ratio, a modulation and coding scheme to be used for the first transmission, an indication of a signal carrier to be used for the first transmission, an information of a physical resource block set allocated for the first transmission, a maximum power reduction to be applied to the first transmission, or a number of transmit layers to be used for the first transmission. The means for selecting may comprise a neural network. The means for selecting may be additionally configured to select a first modulation order to be used for the first transmission a set of plural modulation orders; the means for instructing may be configured to instruct the transmitter to use a modulation of the first modulation order for the first transmission. According to a second aspect, there is provided an apparatus comprising means for selecting, based on a first transmission received by a base station from a terminal and assistance information, a first waveform among a set of plural waveforms; means for demodulating the first transmission assuming that the first waveform is used for the first transmission, wherein the assistance information comprises a channel condition of a channel between the terminal and the base station. The apparatus may further comprise means for inhibiting the base station to instruct the terminal to use the first waveform for the first transmission. The means for selecting may be configured to select the first waveform without having received any information from the terminal that the first waveform is used for the first transmission. The apparatus may further comprise at least one of means for instructing the terminal to take into account at least one parameter of a hysteresis for a switching from the first waveform to a second waveform different from the first waveform; or means for inhibiting the terminal to select any waveform for the first transmission and for instructing the terminal to use a third waveform for the first transmission. The assistance information may comprise a history of waveforms previously used for respective one or more transmissions from the terminal to the base station. The apparatus may further comprise means for receiving an information from the terminal that the waveform used for the first transmission is different from a waveform used for a transmission preceding the first transmission; wherein the means for selecting may be configured to use the information that the first waveform is different from the waveform used for the preceding the first transmission for the selecting the first transmission. The assistance information may comprise at least one of a downlink signal to interference and noise ratio, a modulation and coding scheme to be used for the first transmission, an indication of a signal carrier to be used for the first transmission, an information of a physical resource block set allocated for the first transmission, a maximum power reduction to be applied to the first transmission, or a number of transmit layers to be used for the first transmission. The means for selecting may comprise a neural network. The means for selecting may be additionally configured to select, based on the received first transmission and the assistance information, a first modulation order among a set of plural modulation orders; the means for demodulating may be configured to demodulate the first transmission assuming that a modulation of the first modulation order is used for the first transmission. The apparatus may further comprise means for receiving the first transmission. According to a third aspect, there is provided a method comprising selecting a first waveform among a set of plural waveforms to be used for a first transmission from a terminal to a base station based on assistance information; instructing a transmitter of the terminal to perform the first transmission to the base station using the first waveform; wherein the assistance information comprises a channel condition of a channel between the terminal and the base station. The first waveform may be selected without using an instruction of the waveform to be used received from the base station. The method may further comprise inhibiting the terminal to inform the base station that the first waveform is selected for the first transmission. The method may further comprise receiving an instruction on a hysteresis from the base station; prohibiting the selecting of a second waveform different from the first waveform for a second transmission following the first transmission if a condition defined by the instruction on the hysteresis is not fulfilled. The method may further comprise receiving, from the base station, a prohibition to select the first waveform based on the assistance information and an instruction to use a third waveform for the first transmission; prohibiting the selecting the first waveform based on the assistance information in response to receiving the prohibition; wherein the transmitter of the terminal may be instructed to perform the first transmission using the third waveform in response to receiving the instruction to use the third waveform. The assistance information may comprise a history of waveforms previously used for respective one or more transmissions from the terminal to the base station. The method may further comprise checking whether the first waveform is different from a waveform used for a transmission from the terminal to the base station preceding the first transmission; informing the base station that the first waveform is different from the waveform used for the transmission from the terminal to the base station preceding the first transmission in response to checking that the first waveform is different from the waveform used for the transmission from the terminal to the base station preceding the first transmission. The assistance information may comprise at least one of a downlink signal to interference and noise ratio, a modulation and coding scheme to be used for the first transmission, an indication of a signal carrier to be used for the first transmission, an information of a physical resource block set allocated for the first transmission, a maximum power reduction to be applied to the first transmission, or a number of transmit layers to be used for the first transmission. The selecting may be performed using a neural network. The method may further comprise selecting a first modulation order to be used for the first transmission among a set of plural orders; the transmitter may be instructed to use a modulation of the first modulation order for the first transmission. According to a fourth aspect, there is provided a method comprising selecting, based on a first transmission received by a base station from a terminal and assistance information, a first waveform among a set of plural waveforms; demodulating the first transmission assuming that the first waveform is used for the first transmission, wherein the assistance information comprises a channel condition of a channel between the terminal and the base station. The method may further comprise inhibiting the base station to instruct the terminal to use the first waveform for the first transmission. The first waveform may be selected without having received any information from the terminal that the first waveform is used for the first transmission. The method may further comprise at least one of instructing the terminal to take into account at least one parameter of a hysteresis for a switching from the first waveform to a second waveform different from the first waveform; or inhibiting the terminal to select any waveform for the first transmission and for instructing the terminal to use a third waveform for the first transmission. The assistance information may comprise a history of waveforms previously used for respective one or more transmissions from the terminal to the base station. The method may further comprise receiving an information from the terminal that the waveform used for the first transmission is different from a waveform used for a transmission preceding the first transmission; wherein the information that the first waveform is different from the waveform used for the transmission preceding the first may be used for the selecting the first transmission. The assistance information may comprise at least one of a downlink signal to interference and noise ratio, a modulation and coding scheme to be used for the first transmission, an indication of a signal carrier to be used for the first transmission, an information of a physical resource block set allocated for the first transmission, a maximum power reduction to be applied to the first transmission, or a number of transmit layers to be used for the first transmission. The selecting may be performed using a neural network. Based on the received first transmission and the assistance information, a first modulation order among a set of plural modulation orders may be selected; the first transmission may be demodulated assuming that a modulation of the first modulation order is used for the first transmission. The method may further comprise receiving the first transmission. Each of the methods of the third or fourth aspects may be a method of waveform switching. According to a fifth aspect, there is provided a computer program product comprising a set of instructions which, when executed on an apparatus, is configured to cause the apparatus to carry out the method according to any of the third or fourth aspects. The computer program product may be embodied as a computer-readable medium or directly loadable into a computer. According to some example embodiments, at least one of the following advantages may be achieved: • Additional control signaling for waveform switching is avoided; • Existing DCI signaling may be used. Brief description of the drawings Further details, features, objects, and advantages are apparent from the following detailed description of the preferred example which is to be taken in conjunction with the appended drawings, wherein: Fig.1 illustrates a waveform detector of a gNB including a predictor after OFDM demodulation according to some example embodiments; Fig.2 illustrates a waveform detector of a gNB including a predictor after OFDM demodulation according to some example embodiments; Fig.3 illustrates a waveform detector of a gNB including a predictor after OFDM demodulation according to some example embodiments; Fig.4 illustrates an embedding layer for feeding a waveform index to the ML model according to some example embodiments; Fig.5 illustrates a convolutional neural network architecture for predicting (selecting) the UL waveform at the gNB according to some example embodiments; Fig.6 illustrates a waveform and modulation order detector of a gNB including a predictor after OFDM demodulation according to some example embodiments; Fig.7 illustrates hysteresis parameters; Fig.8 shows a message flow according to some example embodiments; Fig.9 shows an apparatus according to an example embodiment; Fig.10 shows a method according to an example embodiment; Fig.11 shows an apparatus according to an example embodiment; Fig.12 shows a method according to an example embodiment; and Fig.13 shows an apparatus according to an example embodiment. Detailed description of certain example embodiments Herein below, certain example embodiments are described in detail with reference to the accompanying drawings, wherein the features of the example embodiments can be freely combined with each other unless otherwise described. To improve the UL performance, it is recommendable that a 5G NR UE switches dynamically between a single carrier waveform, such as DFT-S-OFDM, and a multi-carrier waveform, such as CP-OFDM, depending on the channel conditions between the UE and the gNB. The usage of the waveforms may be controlled by the gNB (e.g. via an enhanced DCI). Thus, it is ensured that the same waveform is used in the receiver of the gNB to successfully decode the UL signal from the UE. However, the DCI enhancement results in both a reduction in spectral efficiency and a DCI message size because the legacy DCI messages are shorter than the enhanced DCI messages. According to some example embodiments, each of the UE and the gNB may control the waveform for the UL transmission dynamically and autonomously. I.e., in some example embodiments, the gNB does not instruct the UE which waveform to use for an UL transmission from the UE to the gNB. In some example embodiments, the UE does not inform the gNB at all on the waveform used for the UL transmission (“feedback-less”), while in other example embodiments, UE may inform the gNB whether the waveform was switched between subsequent UL transmissions or even about the waveform actually used for the UL transmission. To ensure that gNB uses the same waveform for the demodulating and decoding of the UL transmission as that selected and used by the UE for the UL transmission, the UE selects and the gNB predicts the waveform for the UL transmission based on assistance information which comprises a channel condition of the channel between UE and gNB. At least some of the assistance information is used by both the UE and gNB. Insofar, the selector at the UE and the predictor at the gNB correspond to each other. I.e., the predictor at the UE may be considered as a selector, too. In some example embodiments, if the UE provides some information about the used waveform to the gNB, the predictor of the gNB may use this information in addition to the assistance information used by the UE. In addition, in some example embodiments, at least one of the UE and the gNB may use information on the channel condition as assistance information which is not available to the other one of the UE and the gNB. For example, the UE may use single measurements of a reference signal (e.g. CSI-RS, channel state information reference signal), wherein plural such CSI-RS measurements are compressed for the transmission to the gNB such that the gNB is not aware of the single measurements. On the other hand, gNB may use its measurements of the UL channel, such as an estimated UL signal quality (e.g. SINR), but does not inform UE thereof. In addition to the assistance information, the observed UL signal samples are input into the predictor of the UL waveform in the gNB. When the predictor (selector) of the gNB selects one of the waveforms, it triggers the corresponding UL receiver chain in the gNB for the demodulating and decoding of the UL For the selection / prediction of the UL waveform, the selector / predictor in UE and gNB, respectively, may use at least one of the following pieces of information as assistance information: the channel conditions in downlink (DL) (e.g., channel response information of the physical downlink shared channel (PDSCH) or physical downlink control channel (PDCCH), etc.), the measured DL signal to interference and noise ratio (SINR), the current modulation and coding scheme (MCS), the signal carrier, the set of physical resource blocks (PRB) allocated for the UL transmission, maximum power reduction (MPR), the number of transmit (TX) layers that may be used for the UL transmission, or the history of the last X waveforms used for previous UL transmissions. The UE reports the channel conditions in DL incl. SINR to the gNB such that both UE and gNB are aware of them. The gNB controls the current modulation and coding scheme (MCS), the signal carrier, the set of physical resource blocks (PRB) allocated for the UL transmission, maximum power reduction (MPR), and the number of transmit (TX) layers that may be used for the UL transmission, such that both UE and gNB are aware of them. Each of the selector / predictor in UE and gNB may be a learning tool, e.g. based on a neural network. The prediction (selection) of the UL waveform by the gNB may be seen as a classification problem, wherein the predictor of the gNB outputs a probability of the usage for each UL waveform of the UL waveforms in a candidate set S. For example, in 5G NR the candidate set S may comprise DFT-S-OFDM and CP-OFDM. The prediction of the UL waveform by the gNB according to some example embodiments is described at greater detail: In case the UE transmits the UL signal using CP-OFDM waveform, the gNB receives a signalwhich may be represented as a signal vector ^^ = ^^^^ + ^^ (4) . In case the UEtransmits the UL signal using DFT-S-OFDM, the gNB receives a signal which may berepresented as a signal vector ^^ = ^^^^^ + ^^ (6) . For the details, it is referred to theexplanation further below. Regardless of the TX waveform, each received symbol vector ^^of the frame is then passed through a subcarrier de-mapping block, where the pilot subcarriers are isolated and used forcomputing an estimated channel response ^^ , and ^^ an estimated SINR level ^^^ .Conventionally, the gNB knows the TX because the gNB instructs the UE which waveform to use. I.e., the gNB knows whether the received signal is of the form (4) or (6), so the channel estimates are used to equalize the data symbols which are further demodulated and decoded according to the receive (RX) chains for CP-OFDM or DFT-S-OFDM, to finallyobtain bit estimates ^ ^^, ^ ∈ ^^ − 1^, for all symbols in the frame.In contrast, to some example embodiments, the gNB selects (predicts) first autonomously the waveform, prior to applying the correct (i.e., waveform specific) equalization and demodulation. In other words, the gNB attempts to predict, whether the received signal samples ^^follow the CP-OFDM model according to (4) or the DFT-S-OFDM model according to (6). To that end, the gNB uses the received signal vector ^^, together with assistance information(typically comprising at least the estimated channel response ^^^ and ^^ ) for furtherprocessing. For example, for this prediction, the gNB may use learning. Some implementations of the predictor (selector) in the gNB according to example embodiments are explained at greater detail with reference to Figs. 3 to 5. The predictor (selector) may also be called a deep receiver block. According to Fig. 1, the UL deep receiver block operates on ^^and using assistance information, attempts to classify the transmitted waveform among all the defined waveforms in the candidate set S. The assisting information may be, for example: -DL / UL channel estimates ^^^- SINR 1 / ^^ - Selected MCS - Signal carrier - The allocated PRB set - MPR - The number of TX layers. Note that feeding history of past UL waveforms indices may lead to error propagation in waveform detection results, so it can be considered as an optional input to the waveform detector. Thus, this information is typically not fed into the waveform detector of Fig.1. The predictor (selector) of Fig. 1 classifies the received signals among all the candidates for waveforms. In case the UE decides to switch UL waveform to another UL waveform for subsequent transmissions, according to some example embodiments, the UE may inform gNB that the waveform will be switched. However, due to latency and sparse occasions for the UE to inform the gNB on the switch of the waveform, the gNB does not know exactly which received UL transmission is still transmitted with the previous UL waveform, and which UL transmission is transmitted with the other UL waveform. Nevertheless, the predictor in gNB may take this information into account as another input to select the waveform for each of the UL transmissions. This is shown in Fig.2. In the example embodiment of Fig. 2, the output of the predictor (selector) of the gNB is a decision whether or not the waveform of subsequently received UL transmission has changed. Hence, if the candidate set comprises more than two UL waveforms, the model should be trained for each combination of two UL waveforms of the candidate set. In some example embodiments, such as the one of Fig.3, the UE may inform the gNB not only that the waveform is switched between subsequent UL transmissions, but also on the UL waveform used after the switch, e.g. by a waveform index (WI). However, similarly as explained with respect to Fig. 2, gNB does not know precisely which received UL transmissions were sent using the previous UL waveform and which UL transmission were sent using the other UL waveform. Nevertheless, the predictor (selector) in the gNB may take the information about the other UL waveform into account when selecting a UL waveform for each received UL transmission. The output of the predictor may be again a decision, for each received UL transmission, whether the previous or another UL waveform was used (as shown in Fig.3. As another option (not shown) the predictor may output the waveform index of the likely used UL waveform. Each of the selectors (predictors) in the UE and the gNB may be a Convolutional Neural Network (CNN) comprising convolutional similar to the one shown in Fig. 4. The predictor in the gNB may use the convolutional layers for extracting features from the received signal, followed by dense layers for waveform classification. Assistance information such as estimated SNR of the received signal can be concatenated with the extracted features and fed as input to the dense layers for waveform classification. Fig.5 shows an example of a CNN model for waveform classification by feeding the received signal to an ML model. Alternatively, recurrent neural network architecture or Transformer architecture can be used for ML-based waveform classification block. In some example embodiments, the waveform selector at the UE and the waveform predictor (selector) at the gNB may be trained jointly. For this purpose, they may use known UL bitvectors ^^, ^ ∈ ^^ − 1^, which the UE and gNB agree upon, a loss function as the binary crossentropy between the estimated bit vector at the output of the RX block in the gNB ^^^and the true bit ^^, and (when training is done via simulation) a set of channel models and SINR regimes. In some example embodiments, the waveform selector at the UE and the waveform predictor (selector) at the gNB may be trained separately from each other. In this case, the UE and gNB do not share any dataset with the other entity. The separate training may be performed as follows: For the waveform selection at the UE, supervised or unsupervised learning may be used. In supervised learning, UE collects labeled data (i.e. known transmit bit vectors) from coordinated operations with the gNBs in previous environments or in controlled lab settings. In unsupervised learning (e.g. reinforcement learning (RL)), a reward function may be defined based on the link quality (such as bit error rate (BER), and / or block error rate (BLER), etc.) before and after a waveform switching command. In separate training of the waveform selector at the UE, the selector (predictor) at the gNB does not train its model but works as usual. I.e., gNB runs its receiver (ML based or legacy) Commented [OEB(1]: I see where the confusion comes from. Independent training refers to the fact that when its Independent from the training of the the UEs, gNB may train its selector (predictor, deep waveform detector) using supervised or unsupervised learning approaches. For supervised learning, labeled data may be collected at a time when the UL waveform by a UE is fixed and known by the gNB. In addition, to collect data for challenging situations (periods where waveform switching may happen) in the training dataset, the gNB may process the received signals with all the waveforms in the candidate Set S (exhaustive search over all the possible waveform RX chains). Then the gNB selects the waveform providing the minimum BER or BLER as the label for a received signal. For obtaining the ground truth for the transmitted bits at gNB, gNB may ask the UE to transmit pseudo random bits with a coordinated random seed. In unsupervised learning, the gNB may use a Reinforcement Learning (RL) solution for deep waveform detection. Based on the selected waveform by the RL agent, a reward can be defined as function of the achieved BER or BLER. In other words, high BER / BLER results in large negative reward, while low BER / BLER results in large positive rewards. Whether a BER / BLER is high or low may be defined based on the estimation of SNR of the received signal. Also, the gNB can configure the UE to transmit pseudo random bits with a coordinated random seed, to ease obtaining the ground truth of the transmitted bits. In some example embodiments, not only the waveform but also the modulation order for an UL transmission may be unknown to the gNB. In such example embodiments, the predictor (selector) at gNB may be used to select a pair of a waveform and a modulation order from a candidate set comprising plural pairs of waveform and modulation order. An example of such a predictor (selector) is shown in Fig.6. Both the selector at UE and the predictor (selector) at gNB may use information about the history of the last X waveforms used for the UL transmissions to avoid a ping pong effect associated with switching too often. For this purpose, in some example embodiments, the gNB may control a hysteresis for the waveform switching. For example, as depicted in Fig.7, gNB may configure the UE with one or two of the following hysteresis parameters: offset and delayed time (time to trigger). These hysteresis parameters have the following effects at the UE: The UE must not switch from one waveform to the other if the difference of the probabilities for these waveforms is less than the offset. After a switch from one UL waveform to another UL waveform, the UE must not switch to another waveform for the delayed time. In some example embodiments, one or both of the hysteresis parameters may be predefined in both gNB and UE. The gNB controls the usage of the selector in the UE and the predictor (selector) in the gNB (life cycle management). For example, the gNB may activate or deactivate the selection / prediction of the UL waveform, control one or more of the hysteresis parameters (if they are not predefined), define and potentially activate a fallback solution. If the selector and predictor are learning tools, the gNB may control the training strategy. An example of such control procedures is shown in Fig.8. As shown in Fig.8, the gNB configures in the UE the hysteresis for the switching of the UL waveform. The UE uses the hysteresis parameters in the selection of the UL waveform. For example, if the delayed time after the last waveform switch has not passed, UE will not switch again to another waveform, even if the other waveform appears to be favorable. I.e., during the delayed time, the UE may not even use the waveform selector but simply keep the same UL waveform. Then, the UE performs a UL transmission to gNB using the selected UL waveform. The gNB receives the UL transmission from the UE. The predictor (selector) of the gNB predicts the UL waveform of the received UL transmission. Then, the RX for the predicted UL waveform is used to demodulate and encode the received UL transmission. The gNB may control (manage) the lifecycle of the autonomous UL waveform selection / prediction periodically and / or event based. Some activities belonging to the lifecycle management are activating, deactivating, retraining, setting start / stop times, setting training strategies and defining fallback solution for the UL waveform section. If the autonomous UL waveform selection / prediction is not used, gNB may control the UL waveform used by UE. Fig.9 shows an apparatus according to an example embodiment. The apparatus may be an terminal (such as a UE) or an element thereof. Fig. 10 shows a method according to an example embodiment. The apparatus according to Fig.9 may perform the method of Fig.10 but is not limited to this method. The method of Fig.10 may be performed by the apparatus of Fig.9 but is not limited to being performed by this apparatus. The apparatus comprises means for selecting 110 and means for instructing 120. The means for selecting 110 and means for may be a selecting means and instructing means, respectively. The means for selecting 110 and means for instructing 120 may be a selector and instructor, respectively. The means for selecting 110 and means for instructing 120 may be a selecting processor and instructing processor, respectively. The means for selecting 110 selects a waveform among a set of plural waveforms to be used for a transmission from a terminal to a base station based on assistance information (S110). The assistance information comprises a channel condition of a channel between the terminal and the base station. The means for selecting 110 may select the waveform without using an instruction of the waveform to be used received from the base station. The means for instructing 120 instructs a transmitter of the terminal to perform the transmission to the base station using the waveform selected by the means for selecting 110 (S120). Fig. 11 shows an apparatus according to an example embodiment. The apparatus may be base station (such as a gNB or an eNB) or an element thereof. Fig. 12 shows a method according to an example embodiment. The apparatus according to Fig.11 may perform the method of Fig.12 but is not limited to this method. The method of Fig.11 may be performed by the apparatus of Fig.12 but is not limited to being performed by this apparatus. The apparatus comprises means for selecting 210 and means for demodulating 220. The means for selecting 210 and means for demodulating 220 may be a selecting means and demodulating means, respectively. The means for selecting 210 and means for demodulating 220 may be a selector and demodulator, respectively. The means for selecting 210 and means for demodulating 220 may be a selecting processor and demodulating processor, respectively. The means for selecting 210 selects, based on, a waveform among a set of plural waveforms (S210). The selection is based on assistance information and a transmission received by a base station from a terminal. The assistance information comprises a channel condition of a channel between the terminal and the base station. The means for demodulating 220 demodulates the transmission assuming that the selected waveform is used for the transmission (S220). Fig.13 shows an apparatus according to an example embodiment. The apparatus comprises at least one processor 810, at least one 820 storing instructions that, when executed by the at least one processor 810, cause the apparatus at least to perform the method according to at least one of the following figures and related description: Fig.10 or Fig.12. Some example embodiments are explained with respect to 5G. However, other example embodiments may be employed in other 3GPP generations, such as 4G, 6G, 7G, etc., or in other environments where different waveforms may be used in UL transmissions. Some example embodiments are explained for a candidate set of two UL waveforms: DFT- S-OFDM or CP-OFDM. However, some example embodiments are not limited to these two UL waveforms. Instead, one or both of these UL waveforms may be replaced by one or two other UL waveforms, respectively, such as by LP-OFDMA (linearly precoded OFDM access). Also, the number of UL waveforms is not limited to two. The number of UL waveforms may be any number larger than 1 such as 2 or 3 or 4, etc.. Some example embodiments are explained where the selection of the UL waveform in the UE and / or the prediction (detection) of the UL waveform in the gNB are made using machine learning. However, instead of machine learning, in some example embodiments, a deterministic procedure may be used for the selection and / or prediction. One piece of information may be transmitted in one or plural messages from one entity to another entity. Each of these messages may comprise further (different) pieces of information. Names of network elements, network functions, protocols, and methods are based on current standards, or are current proposals. These names are not limiting. For example, in other versions or other technologies, the names of these network elements and / or network functions and / or protocols and / or methods may be different, as long as they provide a corresponding functionality. The same applies correspondingly to the terminal. If not otherwise stated or otherwise made clear from the context, the statement that two entities are different means that they perform different functions. It does not necessarily mean that they are based on different hardware. That is, each of the entities described in the present description may be based on a different hardware, or some or all of the entities may be based on the same hardware. It does not necessarily mean that they are based on different software. That is, each of the entities described in present description may be based on different software, or some or all of the entities may be based on the same software. Each of the entities described in the present description may be deployed in the cloud. According to the above description, it should thus be apparent that example embodiments provide, for example, an terminal (such as a UE or a MTC device) or an element thereof (which may or may not be actually integrated in the terminal), an apparatus embodying the same, a method for controlling and / or operating the same, and computer program(s) controlling and / or operating the same as well as mediums carrying such computer program(s) and forming computer program product(s). According to the above description, it should thus be apparent that example embodiments provide, for example, a base station (such as a gNB or eNB) or an element thereof (which may or may not be actually integrated in the base station), an apparatus embodying the same, a method for controlling and / or operating the same, and computer program(s) controlling and / or operating the same as well as mediums carrying such computer program(s) and forming computer program product(s). It is to be expressly understood that the description of certain example embodiments is given by way of example only, and that it is by no way intended to be understood as limiting the disclosure to the disclosed details. Moreover, it is to be understood that the apparatus is configured to perform the corresponding method, although in some cases only the apparatus or only the method are described. Implementations of any of the above described blocks, apparatuses, systems, techniques or methods include, as non-limiting examples, implementations as hardware, software, firmware, special purpose circuits or logic, general purpose hardware or controller or other computing devices, or some combination thereof. Each of the entities described in the present description may be embodied in the cloud. It is to be understood that what is described above is what is presently considered the preferred example embodiments. However, it should be noted that the description of the preferred example embodiments is given by way of example only and that various modifications may be made without departing from the scope of the disclosure as defined by the appended claims. The terms “first X” and “second X” include the options that “first X” is the same as “second X” and that “first X” is different from “second otherwise specified. As used herein, “at least one of the following: ” and “at least one of ” and similar wording, where the list of two or more elements are joined by “and” or “or”, mean at least any one of the elements, or at least any two or more of the elements, or at least all the elements. The term "or" refers to a non-exclusive “or” unless otherwise indicated (e.g., use of “or else” or “or in the alternative”). The waveform DFT-S-OFDM is used to increase the UL coverage, allowing the UE to transmit at higher power compared to CP-OFDM, which is used in NR DL and may be used in NR UL. The following definitions and comparison is reproduced verbatim from https: / / uk.mathworks.com / help / comm / ug / scfdma-vs-ofdm.html: “The premise behind DFT-S-OFDM is to use the concept of OFDM, but precode the input signal such that the OFDM output mimics the characteristics of a single-carrier modulated signal within the same transmission bandwidth. Because DFT-S-OFDM uses OFDM as the main transmission modulation scheme, the communications system preserves advantages of OFDM, such as time-frequency user multiplexing, resistance to frequency-selective fading, and frequency-domain equalization, which tends to be easier to implement than time-domain filter-based equalization. The primary advantage of DFT-S-OFDM is a lower peak-to-average power ratio (PAPR) in the transmit signal relative to that found in OFDM. One OFDM symbol contains many sinusoids with different frequencies and phases. The superposition of the many sinusoids occasionally results in constructive interference between sinusoids, which produces a very high peak amplitude relative to the average power of the signal. The problem with high PAPR occurs in both the digital and analog domains. In the digital domain, a high PAPR means that the output signal occasionally exceeds the dynamic range of the digital-to-analog converter (DAC) when the digital signal converts to an analog signal. This exceedance saturates the digital output, a type of distortion called clipping. In the analog domain, the output signal occasionally enters the nonlinear amplification region in the power amplifier of the transmitter. Nonlinear distortion creates undesirable out-of-band emissions in the form of high-order harmonics. Another advantage of DFT-S-OFDM is its against spectral nulls from frequency- selective fading. In OFDM, a null in a subcarrier results in loss of data in that subcarrier. In DFT-S-OFDM, the input signal spreads across subcarriers, so a null in a subcarrier also spreads across the subcarriers (hence the spreading effect in the DFT-s-OFDM name).” For a CP-OFDM UL transmission of B symbols per frame, during the i-th transmission, ^ ∈^^ − 1^, a vector ^^ ∈ !0, 1#$ of bits is encoded by the UE with a code rate R andinterleaved into the vector: *, / 1^,, %^ = &'(( )) + , ... , .( 0) 2 3(1)with entries ((4)) ∈ is denoted the set !> ∈ℕ|0 ≤ > ≤ =#. The complex symbols that areinterleaved 7B pilots to yield the symbol ^^ ∈ ℂ / , 7B + 78 = D.The m-th entry E)^F^ of ^^ is a pilot symbol if F ∈ = and a data symbol if F ∈ :. Here, = =!>*, ... , > / G1^# and : = !H*, ... , H / 01^#, = ∪ : = ^D − 1^, = ∩ : = ∅, are the subsets =) ,PQ denotes the complex modulation mapping, and PQ is the complex cardinality25. The vector ^^is passed through an inverse discrete Fourier transform (DFT) block and the resulting vector R^is prepended a S-sample long cyclic prefix (CP). Each symbol R^in the B-long symbols frame is sent then by the UE over an UL channel with response composed of L multipath components: [1^ U= V Y.U − 2whereℎ)(X)andU)(X) component. ^ The gNB observes a signal which is at a rate,̂, and the cyclic prefix (CP) samples are discarded. The remaining samples are collected in B vectors that are passed through an M-point DFT block, at the output of which the received vector is: ^^ = ^^^^ + ^^ (4)where ^ ∈ ^^ − 1^ and the matrix ^^ ∈ ℂ_×_ is diagonal, with diagonal entries:^^^F, F^ = ^∑[\]1*^ℎ)(X) exp 'e fghi(\) + with. The vector ^^is additive white Gaussian ^ j / . On the other hand, for a DFT-S-OFDM UL transmission, the UE processing chain applies thesame operations (1)-(2) as above to yield the 7 (l)8 sub-symbols E)^k^ = ℳ'() +, k ∈ ^78 −1^. These are passed through a DFT block, the output of which 1 / 01^−p2q6k Em)(H4) = VE)^k^ exp o 7 r (5)subcarriers set. Then, theth entry Em)^F^ is a pilot ... , > / G1^# and : =!H*, ... , H / 01^#, = ∪ : = ^D − 1^, = ∩ : = ∅, are the subsets of pilot and data indexes,respectively, the same as in the OFDM chain. Finally, the resulting vector is converted to time-domain via inverse DFT of size M, the same as in the CP-OFDM transmission. The resulting vector R^is prepended a S-sample long cyclic prefix (CP). Each symbol R^in the B-long symbols frame is sent then by the UE over an UL channel withresponse composed of L multipath components T [1^)( U ) = ∑\]* ℎ)(X)Y.U − U)(X)2, here ℎ)(X)and U)(X) model the gain and delay of the X-th ^ The gNB observes a signal which is sampled at a rate,̂, and the CP samples are discarded. The remaining samples are collected in B vectors that are passed through an M-point DFT block, at the output of which the received vector is: ^^ = ^^^^^ + ^^ (6)where ^ ∈ ^^ − 1^ and the matrix ^^ ∈ ℂ_×_ is diagonal, with diagonal entries:^^^F, F^ = ^_ ∑[\]1*^ℎ)(X) exp 'e fghi(\)_,̂ +. vector ^ is AWGN w ^ith zero mean and variance ^ j / .

Claims

Claims:

1. Apparatus comprising means for selecting a first waveform among a set of plural waveforms to be used for a first transmission from a terminal to a base station based on assistance information; means for instructing a transmitter of the terminal to perform the first transmission to the base station using the first waveform; wherein the assistance information comprises a channel condition of a channel between the terminal and the base station. 2 The apparatus according to claim 1, wherein the means for selecting is configured to select the first waveform without using an instruction of the waveform to be used received from the base station. 3 The apparatus according to any of claims 1 and 2, further comprising means for inhibiting the terminal to inform the base station that the first waveform is selected for the first transmission. 4 The apparatus according to any of claims 1 to 3, further comprising means for receiving an instruction on a hysteresis from the base station; means for prohibiting the means for selecting to select a second waveform different from the first waveform for a second transmission following the first transmission if a condition defined by the instruction on the hysteresis is not fulfilled. 5 The apparatus according to any of claims 1 to 4, further comprising means for receiving, from the base station, a prohibition to select the first waveform based on the assistance information and an instruction to use a third waveform for the first transmission; means for prohibiting the means for selecting from selecting the first waveform based on the assistance information in response to receiving the prohibition; wherein the means for instructing is configured to instruct the transmitter of the terminal to perform the first transmission using the third waveform in response to receiving the instruction to use the third waveform.

6. The apparatus according to any of claims 1 to 5, wherein the assistance information comprises at least one of a downlink signalinterference and noise ratio, a modulation and coding scheme to be used for the first transmission, an indication of a signal carrier to be used for the first transmission, an information of a physical resource block set allocated for the first transmission, a maximum power reduction to be applied to the first transmission, a number of transmit layers to be used for the first transmission, or a history of waveforms previously used for respective one or more transmissions from the terminal to the base station. 7 The apparatus according to claim 6, further comprising means for checking whether the first waveform is different from a waveform used for a transmission from the terminal to the base station preceding the first transmission; means for informing the base station that the first waveform is different from the waveform used for the transmission from the terminal to the base station preceding the first transmission in response to checking that the first waveform is different from the waveform used for the transmission from the terminal to the base station preceding the first transmission, wherein the assistance information comprises the history of waveforms previously used for respective one or more transmissions from the terminal to the base station. 8 The apparatus according to any of claims 1 to 7, wherein the means for selecting is additionally configured to select a first modulation order to be used for the first transmission among a set of plural modulation orders; the means for instructing is configured to instruct the transmitter to use a modulation of the first modulation order for the first transmission. 9 Apparatus comprising means for selecting, based on a first transmission received by a base station from a terminal and assistance information, a first waveform among a set of plural waveforms; means for demodulating the first transmission assuming that the first waveform is used for the first transmission, wherein the assistance information comprises a channel condition of a channel between the terminal and the base station. 10 The apparatus according to claim 9, further comprisingmeans for inhibiting the base station to instruct the terminal to use the first waveform for the first transmission.

11. The apparatus according to any of claims 9 and 10, wherein the means for selecting is configured to select the first waveform without having received any information from the terminal that the first waveform is used for the first transmission.

12. The apparatus according to any of claims 9 to 11, further comprising at least one of means for instructing the terminal to take into account at least one parameter of a hysteresis for a switching from the first waveform to a second waveform different from the first waveform; or means for inhibiting the terminal to select any waveform for the first transmission and for instructing the terminal to use a third waveform for the first transmission.

13. The apparatus according to any of claims 9 to 12, wherein the assistance information comprises at least one of a downlink signal to interference and noise ratio, a modulation and coding scheme to be used for the first transmission, an indication of a signal carrier to be used for the first transmission, an information of a physical resource block set allocated for the first transmission, a maximum power reduction to be applied to the first transmission, a number of transmit layers to be used for the first transmission, or a history of waveforms previously used for respective one or more transmissions from the terminal to the base station.

14. The apparatus according to claim 13, further comprising means for receiving an information from the terminal that the waveform used for the first transmission is different from a waveform used for a transmission preceding the first transmission; wherein the means for selecting is configured to use the information that the first waveform is different from the waveform used for the transmission preceding the first transmission for the selecting the first transmission, wherein the assistance information comprises the history of waveforms previously used for respective one or more transmissions from the terminal to the base station.

15. The apparatus according to any of claims 9 to 14, whereinthe means for selecting is additionally configured to select, based on the received first transmission and the assistancea first modulation order among a set of plural modulation orders; the means for demodulating is configured to demodulate the first transmission assuming that a modulation of the first modulation order is used for the first transmission.

16. The apparatus according to any of claims 9 to 15, further comprising means for receiving the first transmission.

17. The apparatus according to any of claims 1 to 16, wherein the means for selecting comprises a neural network.

18. Method comprising selecting a first waveform among a set of plural waveforms to be used for a first transmission from a terminal to a base station based on assistance information; instructing a transmitter of the terminal to perform the first transmission to the base station using the first waveform; wherein the assistance information comprises a channel condition of a channel between the terminal and the base station.

19. Method comprising selecting, based on a first transmission received by a base station from a terminal and assistance information, a first waveform among a set of plural waveforms; demodulating the first transmission assuming that the first waveform is used for the first transmission, wherein the assistance information comprises a channel condition of a channel between the terminal and the base station.

20. A computer program product comprising a set of instructions which, when executed on an apparatus, is configured to cause the apparatus to carry out the method according to any of claims 18 or 19.

Images