Determination of a custom constellation based on a calibration signal

AI-based custom constellations with I-Q-transformations address the challenge of adapting modulation schemes to individual radio and hardware conditions, improving spectral efficiency and RF compliance in communication networks.

WO2026233672A1PCT designated stage Publication Date: 2026-11-12NOKIA TECHNOLOGIES OY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
PCT/EP2026/061202
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2025-05-08
Filing Date
2026-04-17
Publication Date
2026-11-12

AI Technical Summary

Technical Problem

Current communication networks struggle to adapt modulation schemes dynamically to individual radio environments and hardware constraints, particularly in handling non-linear power amplifier distortions, leading to non-compliance with radio frequency requirements.

Method used

Implementing AI-based custom constellations with I-Q-transformations to adjust constellation points in real-time, optimizing for both out-of-band and in-band RF requirements by determining a constellation distortion array based on RF parameters.

Benefits of technology

This approach enhances spectral efficiency and compliance with RF requirements by minimizing distortions caused by non-linear components while maintaining acceptable bit error rates.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure EP2026061202_12112026_PF_FP_ABST
    Figure EP2026061202_12112026_PF_FP_ABST
Patent Text Reader

Abstract

Training and transferring a constellation distortion array, CDA, between a network node and a terminal device. The terminal device indicates a capability to support custom constellation training (S201). The network node requests a known calibration signal (S202), which the terminal device transmits (S203). The network node uses the received calibration signal to model a power amplifier response of the terminal device and derives a CDA configured to minimize radio frequency requirement violations (S204- S207). The network node transfers the derived CDA to the terminal device (S208), which applies the CDA to transform a reference constellation into a custom constellation for uplink transmission (S209).
Need to check novelty before this filing date? Find Prior Art

Description

[0001] Constellation adjustment

[0002] TECHNOLOGICAL FIELD

[0003] The present disclosure is related but not limited to communication networks as defined by the 3GPP standard, such as the 5G and / or 6G standard. The disclosure in particular relates to modulation, in particular to adjusting constellations used in modulation, in particular for uplink communication from a terminal device to a network node.

[0004] BACKGROUND

[0005] In mobile communication networks, such as those complying with 3GPP standards, modulation schemes (e.g., techniques for converting digital values to physical radio frequency waveforms, for instance, by adjusting the amplitude and / or phase of carriers / subcarriers] maybe selected and / or adjusted according to pre-defined mechanisms. Such selection and / or adjustment may be performed for uplink and / or downlink transmissions and may for instance have the goal to ensure at least one of spectral efficiency, acceptable error rates, interoperability, adaptation to radio conditions, and / or robustness against hardware impairments. One of such mechanisms is a selection of a modulation and coding schemes (MCS] by a network node by means of indexing an MCS table storing different MCS to select. The MCS table may define a structured approach for adjusting transmission parameters to different radio link conditions. This may allow to adapt modulation formats (e.g., number of constellation points in a modulated representation] dynamically based on a predefined set of MCS to select from.

[0006] SUMMARY OF SOME EXEMPLARY EMBODIMENTS

[0007] It has been recognized that currently used methods of adapting modulation and coding schemes and in particular modulation schemes are based on a utilization of pre-defined, invariant modulation schemes.

[0008] A modulation may in particular comprise a constellation, wherein the constellation comprises constellation points within an in-phase and quadrature, I-Q, plane. One constellation point may be associated to a binary symbol or binary label. A binary symbol is a binary string with a given length (e.g., number of bits], E.g., all binary symbols of a constellation (e.g., associated to respective constellation points] may have a same length. The length of the binary symbols may match a modulation order of the constellation (e.g., m; e.g., for a constellation of order m there may be 2mconstellation points, a respective one associated to a unique binary symbol]. The binary symbols are associated with respective constellation points according to a mapping rule, such as Gray mapping. For instance, a constellation point (i=l, q=l] may be mapped to a binary symbol of ‘00’ in a 4-QAM constellation. A constellation point corresponding to a given binary symbol is then selected to yield a modulation symbol for a respective sub-carrier. The modulation symbol may correspond to a complex coefficient derived from the selected constellation point (e.g., corresponding to a respective binary symbol], and may further be input to an inverse Fast Fourier Transform, IFFT, block (possibly after precoding or some other kind of signal pre-processing] to yield an OFDM symbol (after cyclic extension].For instance, such modulation schemes may comprise Quadrature amplitude modulation [QAM] or Quadrature Phase Shift Keying (QPSK], e.g. of different sizes (e.g., different numbers of constellation points, e.g., one per binary symbol]. Such pre-defined modulation schemes typically comprise regular and / or fixed positions of constellation points within an in-phase and quadrature, I-Q, plane. Current approaches may therefore be limited to adjust a type and / or a size (e.g., constellation point count] of a pre-defined modulation scheme. Such approaches are for instance defined in TS 38.214 and TS 38.211, wherein M-QAM modulations are used for Physical Uplink SHared Channel, PUSCH, and Physical Downlink SHared Channel, PDSCH. A network node, e.g., gNB, may select a modulation and coding scheme [MCS] by indexing an MCS table based on Channel Quality Indicator (CQI] reports from the UE, thereby ensuring adaptive modulation. Higher CQI values enable higher-order constellations.

[0009] It has further been recognized that future transmissions in mobile communication networks may not be limited to such pre-defined modulation schemes but may instead implement custom constellations that depart form the currently used pre-defined, regular, and / or fixed modulation schemes. Such custom constellations may provide more fine-grained and / or individual adjustments of modulation schemes and / or constellation point arrangement to conditions such as a current radio environment and / or hardware constraints. Future modulation schemes may even be individually optimized to individual radio communication links between a given terminal device and a given network node, e.g., at a given point in time and space.

[0010] For instance, custom constellations may correspond to artificial intelligence, Al, based learned constellations. For instance, a custom constellation may be obtained by jointly optimizing symbol mapping and detection, e.g., through neural networks, in particular autoencoder models. Unlike current modulation schemes such as QAM, which follow a rigid (e.g., pre-defined, invariable] structure (e.g., constellation point arrangement], AI-based modulation schemes may allow adaptation of constellation point arrangements to channel conditions and / or hardware impairments (e.g., non-linear power amplifiers]. Learned constellations have shown promise to at least one of (1] Reduce pilot overhead, e.g., as neural receivers may be able to jointly perform channel estimation and symbol detection, (2] adapt to power amplifier, PA, distortions, e.g., by positioning constellation points to be more resilient to nonlinear effects, and / or (3] outperform standard QAM, by achieving comparable or better bit error rates while improving spectral efficiency.

[0011] As a result, a multitude of mutually different and potentially irregular custom constellations may thus be employed in future communication networks. It has been recognized that current approaches of (e.g., dynamic] adjustment of modulation schemes (e.g., constellations] to (e.g., radio] conditions such as an adjustment of constellation point count by means of a MCS tables cannot be straightforwardly and / or efficiently reused for such custom constellations.

[0012] It has further been recognized that current approaches may not sufficiently account for hardware conditions at the terminal device such as nonlinearities of power amplifiers, PA. PA nonlinearity may for instance becaused by memory effects which may for instance cause dynamic distortion patterns, e.g., dependent on signal characteristics and operation conditions. PA nonlinearities may introduce amplitude and / or phase distortions that affect signal quality. Currently employed, fixed constellations (e.g., QAM] may be susceptible to such hardware conditions to (e.g., only] a moderate extent. Current communication networks may therefore mitigate the effects of such PA nonlinearities through general measures such as Discrete Fourier Transform-Spread-Orthogonal Frequency Division Multiplexing, DFT-S-OFDM, for uplink (e.g., reducing Peak to Average Power Ratio, PAPR] and / or TT / 2-BPSK, e.g., for extreme coverage. Such counter-measures do, however, not dynamically adapt to (e.g., individual] terminal device characteristics but merely reduce the effects of (e.g., unknown] nonlinearities.

[0013] It has further been recognized that (e.g., inherent] nonlinear behavior of power amplifiers (PAs] may lead to spectral regrowth, introducing out-of-band emissions (e.g., measured as Adjacent Channel Leakage Ratio, ACLR] and intermodulation products (e.g., measured as Error Vector Magnitude, EVM], The OFDM transmitter is known to introduce nonlinear distortions to the transmitted signal due to hardware imperfections. The power amplifier, as a major source of nonlinear distortion, is known to create power leakage, both in band and out of band. Without counter measures, a terminal device may thus violate radio frequency requirements.

[0014] Radio frequency, RF, requirements may for instance comprise at least one of the following (e.g., according to TS 38.101], (1] Spectrum emission mask (SEM]: The SEM of the UE applies to frequencies (out-of-band frequencies, Afoos] starting from the edge of a channel bandwidth assigned to the UE. Separate requirement tables may apply for Cyclic Prefix Orthogonal Frequency Division Multiplexing, CP-OFDM, and Discrete Fourier Transform Spread OFDM, DFT-S-OFDM. SEM may be UE-unspecific. (2] Adjacent Channel Leakage power Ratio (ACLR] may be defined as a ratio of a filtered mean power centered on the assigned channel frequency to a filtered mean power centered on an adjacent channel frequency. Different requirements may be defined for mutually different power classes. Different requirements may be defined when power boosting is enabled, e.g., for Binary Phase Shift Keying, BPSK, and / or Quadrature Phase Shift Keying, QPSK. (3] In band emissions may for instance relate to a ratio of the UE output power in a non-allocated Resource Block, RB, to a UE output power in an allocated RB. (4] Occupied BW may for instance be defined as the bandwidth containing 99 % of a total integrated mean power of a transmitted spectrum on the assigned channel. (5] EVM may for instance (e.g., for DFT-s-OFDM] be defined for a point in a signal chain after a front-end Fast Fourier Transform, FFT, and Inverse Discrete Fourier Transform, IDFT, have been performed. Additionally or alternatively, e.g., for CP-OFDM, EVM may be defined for a point in a signal chain after the front-end FFT. EVM may be computed as a square root of a ratio of the mean error vector power to the mean reference power, e.g., expressed as a percentage. (6] Transmit intermodulation may for instance relate to a measure of a capability of the transmitter to inhibit a generation of signals in its non-linear elements caused by presence of the wanted signal and an interfering signal reaching the transmitter via the antenna.

[0015] As one approach to handle nonlinearities at the power amplifier and, inter alia, conform with RF requirements, Digital Pre Distortion, DPD. For instance, so called memory polynomial-based DPD relies onmodeling and estimating PA characteristics at the transmitter. DPD approaches typically require feedback loops to measure the output and compute predistortion parameters. A common drawback of DPD solutions is their reliance on a wide feedback bandwidth, often requiring up to three times the fundamental frequency to capture sufficient relevant intermodulation products. This feedback requirement significantly increases complexity and hardware costs. Moreover, DPD systems are online systems, so that each transmission needs to recalculate DPD signals. DPD is sub-optimal in view of desired (e.g., in 6g] always available principles. Conventional DPD methods thus aim to counteract PA distortions, but their performance is limited in adaptive or complex scenarios.

[0016] As a further approach, Al receivers, particularly neural receivers, may prove able to estimate and / or recover transmitted signals in complex communication environments. Neural networks may be used to jointly optimize signal processing at both the transmitter and receiver, incorporating constraints such as adjacent channel leakage ratio (ACLR] and peak-to-average power ratio (PAPR] to ensure compliance with hardware and spectral limitations. Al receivers may leverage data-driven training on channel models, focusing on static or semi-static configurations, in order to outperform traditional model-based receivers in terms of bit error rate (BER] and adaptability to non-linear impairments. Al receivers acting in time and frequency domains may further alleviate power amplifier [PA] nonlinearities. Applications of Al transceivers in digital postdistortion and signal enhancement may thus hold promise not only for enhanced spectral efficiency but also for achieving compliance with radio frequency requirements, e.g., emission requirements. Yet, such Al systems require either joint training of transmitter and receiver or prior knowledge of the PA non-linearity, which can limit their practicality in wireless environments where prior knowledge can be difficult to obtain.

[0017] Further, it has been recognized that Al receivers implementing digital post distortion mechanisms have been considered to compensate for the in-band nonlinearities. However, the uplink transmission emissions that pollute parts of the spectrum which are adjacent to a UE-assigned bandwidth have so far not been mitigated.

[0018] In summary, there currently exists a lack of real-time compensation mechanisms that allow to conform with RF requirements in view of nonlinear power amplifier behavior.

[0019] It is one of the objects of the present disclosure to provide apparatus and methods for complying particularly well with RF requirements, in particular for uplink transmission (e.g., in view of power amplifier nonlinearities], in particular when utilizing custom constellations.

[0020] It has been recognized that RF requirements may be complied with by leveraging transformations of constellations in I-Q-plane, which may be referred to by I-Q-transformation. An I-Q-transformation shapes a given constellation to dynamically adjust the arrangement of constellation points (e.g., corresponding to binary symbols] to achieve specific performance goals. Such performance goals may in particular relate to RF requirements such as, for instance, reducing peak-to-average power ratio (PAPR] and / or optimizing spectral efficiency. It is one of the advantages of custom constellations (e.g., over pre-defined and / or regular modulation schemes, e.g., square modulation schemes like QAM] to be able to leverage adaptive techniquessuch as I-Q-transformation to minimize distortions caused by non-linear components while maintaining acceptable bit error rate (BER] performance and / or complying with RF requirements. For instance, Al models (e.g., neural networks] may be used to derive I-Q-transformations to optimize at least one RF requirement (e.g., reduced PAPR], e.g., while maintaining BER. It has been recognized that combining Al models with custom constellations, in particular with adjusting constellations by means for I-Q-transformation, may be an effective method for real-time performance improvements. It has been recognized that a transmitter may for instance transform a transmitted signal in such a way that a resulting waveform complies with a preselected out-of-band RF requirements (e.g., SEM and ACLR limits]. Yet, such transformations may worsen in-band leakage so that a challenge may be to determine transformations that make the UE comply with both out-of-band and in-band RF requirements (e.g., leakage], in particular in view of (e.g., unknown to the network] power amplifier non-linearities.

[0021] According to a first example aspect, a method is disclosed, for instance performed by a terminal device (e.g., user equipment, UE], comprising:

[0022] obtaining at least one radio frequency, RF, requirement parameter from a network node; obtaining a constellation request for a custom constellation from the network node; determining a constellation distortion array (e.g., a matrix, e.g. of dimension 2xN] based on the at least one RF requirement parameter, wherein a respective entry of the constellation distortion array indicates a shift in an In-phase / Quadrature, I / Q, plane of a respective reference constellation point of a reference constellation into a corresponding custom constellation point of the custom constellation; and

[0023] providing an indication of the determined constellation distortion array to the network node.

[0024] According to a second example aspect, a method is disclosed, for instance performed by a terminal device (e.g., user equipment, UE], comprising:

[0025] providing (e.g., transmitting] a calibration signal to a network node using a reference constellation; and

[0026] obtaining an indication of a constellation distortion array (e.g., based on the calibration signal] from the network node in response to providing the calibration signal, wherein a respective entry of the constellation distortion array indicates a shift in an In-phase / Quadrature, I / Q, plane of a respective reference constellation point of the reference constellation into a corresponding custom constellation point of a custom constellation.

[0027] According to a third example aspect, a method is disclosed, for instance performed by a terminal device (e.g., user equipment, UE], comprising:

[0028] obtaining (e.g., an indication of] a constellation distortion array, wherein a respective entry of the constellation distortion array indicates a shift in an In-phase / Quadrature, I / Q, plane of a respective reference constellation point of a reference constellation into a corresponding custom constellation point of a custom constellation;modulating uplink information using the reference constellation (e.g., QAM] into at least one reference constellation point of the I / Q plane;

[0029] determining a shift to be applied to the reference constellation point based on the constellation distortion array;

[0030] shifting the at least one reference constellation point according to the determined shift into at least one respective custom constellation point of the custom constellation; and

[0031] performing an uplink transmission (e.g., PUSCH] to a network node, wherein the uplink transmission comprises the at least one custom constellation point.

[0032] The method according to the first, second or third example aspect may for instance be performed and / or controlled by an apparatus, for instance a server. Alternatively, the method may be performed and / or controlled by more than one apparatus, for instance a server cloud comprising at least two servers.

[0033] Alternatively, the method may for instance be performed and / or controlled by an electronic device, e.g. a node in a communication system and / or by a terminal device, e.g., a user equipment (UE], For instance, the method may be performed and / or controlled by using at least one processor of the electronic device.

[0034] According to a further example aspect, a computer program is disclosed, the computer program when executed by a processor causing an apparatus, for instance a server, a network node or a terminal device, e.g., a UE, to perform and / or control the actions of the method according to the first, second or third example aspect.

[0035] The computer program may be stored on computer-readable storage medium, in particular a tangible and / or non-transitory medium. The computer readable storage medium could for example be a disk or a memory or the like. The computer program could be stored in the computer readable storage medium in the form of instructions encoding the computer-readable storage medium. The computer readable storage medium may be intended for taking part in the operation of a device, like an internal or external memory, for instance a Read-(e.g., Only] Memory [ROM] or hard disk of a computer, or be intended for distribution of the program, like an optical disc.

[0036] According to a further aspect, a computer program product is disclosed comprising the computer program.

[0037] According to a further example aspect, an apparatus is disclosed, configured to perform and / or control or comprising respective means for performing and / or controlling the method according to the first, second or third example aspect. The means may for instance comprise a power amplifier, for instance in a transmission chain for uplink transmissions to a network node.

[0038] The means of the apparatus can be implemented in hardware and / or software. They may comprise for instance at least one processor for executing computer program code for performing the required functions, at least one memory storing the program code, or both. Alternatively, they could comprise for instance circuitry that is designed to implement the required functions, for instance implemented in a chipset or achip, like an integrated circuit. In general, the means may comprise for instance one or more processing means or processors.

[0039] The above-disclosed apparatus according to any aspect may be a module or a component for a device, for example a chip. Alternatively, the disclosed apparatus according to any aspect may be a device, for instance a server or server cloud. The disclosed apparatus according to any aspect may comprise (e.g., only] the disclosed components, for instance means, processor, memory, or may further comprise one or more additional components.

[0040] According to a further example aspect, an apparatus, e.g., a UE, is disclosed, comprising respective means for performing the actions of the method of any of the first, second or third example aspect, for instance according to at least one or more embodiments of the first, second or third example aspect.

[0041] According to a further example aspect, an apparatus, e.g., a UE, is disclosed, comprising at least one processor; and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to perform and / or control the actions of the method according any of embodiments of any of the first, second or third example aspect, for instance according to at least one or more embodiments of the first, second or third example aspect.

[0042] A terminal device, e.g., a user equipment [UE] may for instance correspond to a mobile device such as, for example, a mobile phone, tablet, smartwatch, a laptop, a Personal Digital Assistant [PDA] device, a wearable, an Internet-of-Things (I0T) device, an IIOT (Industrial I0T] device, a vehicle and / or combinations thereof. Such a user equipment may also be referred to as user device. The terminal device may for instance correspond to and / or comprise a Low-Power Wide-Area [LPWA] device and / or an Enhanced Mobile Broadband [eMBB] device. For instance, the terminal device may comprise at least one or (e.g., or both of) an LPWA module or eMBB module.

[0043] A network node may correspond to a component of a communication network such as for instance a Base Transceiver Station (BTS), a nodeB, an evolved node B (eNB), a Next Generation NodeB (gNB), a distributed unit (DU), a central unit (CU) and / or combinations thereof.

[0044] An obtaining of a quantity (e.g., the RF requirement parameter, the constellation request, the indication of a constellation distortion array and / or the constellation distortion array) by the apparatus performing and / or controlling the method according to the first, second and / or third example aspect, e.g., from the network node, may comprise and / or correspond to receiving the quantity or an according indication of the quantity from the network node. The receiving may be performed by means of at least one of a Downlink Control Information, DCI, a Medium Access Control Control Element, MAC CE, a Radio Resource Control, RRC, signaling and / or combinations thereof.

[0045] A providing of a quantity (e.g., the indication of the determined constellation distortion array) by the apparatus performing and / or controlling the method according to the first, second and / or third exampleaspect to the network node may comprise and / or correspond to transmitting the quantity or an according indication of the quantity to the network node. The transmitting may be performed by means of at least one of an Uplink Control Information, UCI, a Medium Access Control Control Element, MAC CE, a Radio Resource Control, RRC, signaling and / or combinations thereof.

[0046] According to the first example aspect, the method comprises obtaining at least one radio frequency, RF, requirement parameter from a network node.

[0047] The RF requirement parameter may for instance correspond to at least one of an in-band leakage requirement, an out-of-band leakage requirement or both. For instance, the RF requirement may correspond to at least one of Adjacent Channel Leakage Ratio, ACLR, Peak-to-Average Power Ratio, PAPR, Spectrum Emission Mask, SEM, in-band emissions, Error Vector Magnitude, EVM, transmit intermodulation and / or combinations thereof. The RF requirement may indicate a single quantity. Alternatively, the RF requirement may indicate two or more quantities, for instance, at least or at most 2, 3, 4, or 5 quantities. The RF requirement parameter may for instance indicate a threshold for a respective quantity, e.g., to conform with by the apparatus performing and / or controlling the method according to the first, second and / or third example aspect in one of more future uplink transmissions (e.g., one or more Physical Uplink SHared Channel, PUSCH, transmissions].

[0048] According to the first example aspect, the method further comprises obtaining a constellation request for a custom constellation from the network node. The constellation request may for instance be obtained jointly with the RF requirement parameter or separately from the RF requirement parameter. Additionally or alternatively, the constellation request may for instance be obtained with a same signaling type (e.g., DCI, MAC CE or RRC signaling] as the RF requirement parameter or with another signaling type compared to the RF requirement parameter.

[0049] The constellation request indicates a request for a custom constellation. The network node may thus request a custom constellation from the apparatus performing and / or controlling the method according to the first, second and / or third example aspect

[0050] The custom constellation may be different from pre-defined, fixed constellations such as for instance QAM or QPSK. The custom constellation may for instance comprise irregularly spaced constellation points in the I-Q-plane, for instance wherein a first, second or third distance between a given constellation point and a corresponding nearest neighbor constellation point of the given constellation point is different from at least one second distance between another constellation point and a corresponding nearest neighbor constellation point of the other constellation point of the custom constellation. The custom constellation may for instance be specific to the apparatus performing and / or controlling the method according to the first, second or third example aspect, for instance to a characteristic (e.g., non-linearity] of a power amplifier of the apparatus performing and / or controlling the method according to the first, second or third example aspect and / or to a radio communication link between the apparatus performing and / or controlling the method according to thefirst, second or third example aspect and the network node. For instance, the custom constellation may be a learned constellation, e.g., obtained by means of a machine learning model, e.g., by the network node, e.g., based on calibration signals provided by the apparatus performing and / or controlling the method according to the first, second or third example aspect to the network node.

[0051] The constellation request may for instance indicate a request for a custom constellation that causes the apparatus performing and / or controlling the method according to the first, second and / or third example aspect to comply with the RF requirement parameter.

[0052] According to the first example aspect, the method further comprises determining a constellation distortion array based on the at least one RF requirement parameter.

[0053] The constellation distortion array may be indicative of a transformation of a constellation (e.g., a custom constellation or a regular, e.g., legacy, constellation such as QAM] in I-Q-plane. The constellation distortion array may comprise multiple entries, wherein a respective entry of the constellation distortion array indicates a shift in the In-phase / Quadrature, I-Q-plane of a respective reference constellation point of a constellation (e.g., reference constellation] into a corresponding custom constellation point of the custom constellation. A respective shift may for instance correspond to a difference vector between a to be achieved custom constellation point and a corresponding reference constellation point. By means of the constellation distortion array, the reference constellation may be transformed to the custom constellation, e.g., wherein the reference and the custom constellation have a same number of constellation points.

[0054] The constellation distortion array may for instance be represented as a matrix, e.g. a matrix with N rows and 2 columns, wherein a respective pair of two entries in a row indicates a respective shift to be applied to one or more constellation points. A row index may be indicative of an identity of a respective constellation point of the reference constellation. Additionally or alternatively, the constellation distortion array may be encoded as a vector, as a list, as parameters to a pre-defined generating function used to generate the constellation distortion array in a pre-defined (e.g., deterministic] manner, as a dictionary and / or combinations thereof.

[0055] The constellation distortion array may indicate as many shifts as there are constellation points in the reference constellation. One respective shift may be (e.g., only] applied to one respective constellation point. E.g., for a 64 QAM constellation, there may be 64 shifts indicated by the constellation distortion array which may for instance be represented as a 64x2 matrix. Alternatively, the constellation distortion array may indicate less shifts than there are constellation points in the reference matrix, for instance half as many or quarter as many. One respective shift indicated by the constellation distortion array may in this case be applied to more than one (e.g., 2 or 4 in the case of half or quarter as many shifts compared to constellation points] constellation point. It has been recognized that a symmetric constellation may enable an application of a single shift to multiple reference constellation points. For instance, a shift may be applied to 2 reference constellation points in two mutually different halfs of the I-Q-plane (e.g., once to a constellation point in the half Q>=0 and once to a constellation point in the half Q<0; e.g., once to a constellation point in the half I>=0and once to a constellation point in the half I<0], For instance, a shift may be applied to 4 reference constellation points in four mutually different quadrants (e.g., once to a constellation point in the quadrant Q>=0, I>=0 and once to a constellation point in the quadrant Q<0, I>=0 and once to a constellation point in the half Q>=0, I<0 and once to a constellation point in the half Q<0, 1 < 0 ] . For instance, the respective shift may undergo at least one of a mirroring about the I- and / or Q-axis when being applied to reference constellation points in mutually different halves or quadrants of the I-Q-plane. For instance, a shift may be applied as indicated by the constellation distortion array in a first quadrant Q>=0, I>=0 but be mirrored about the I-axis when applied in the quadrant Q>=0, 1 <0, mirrored about the Q-axis when applied in the quadrant Q<0, I>=0 and mirrored about both I- and Q-axis when applied to the quadrant Q<0, I<0.

[0056] The constellation distortion array may be based on the RF requirement parameter in that the apparatus performing and / or controlling the method according to the first, second and / or third example aspect (e.g., is expected to fulfil and / or] fulfils the RF requirement parameter when using the custom constellation obtained by means of the constellation distortion array for uplink transmissions. Additionally or alternatively, the obtaining of the constellation distortion array may be based on the RF requirement parameter in that the RF requirement parameter is used to determine (e.g., select, generate, compute and / or derive] the constellation distortion array.

[0057] The constellation distortion array may for instance be determined based on a characteristic of a power amplifier of the apparatus performing and / or controlling the method according to the first, second and / or third example aspect, e.g., wherein the characteristic of the power amplifier is known to the apparatus performing and / or controlling the method according to the first, second and / or third example aspect. For instance, the apparatus performing and / or controlling the method according to the first, second and / or third example aspect may thus (e.g., through knowing the power amplifier characteristics] be enabled to select a correct custom constellation and / or corresponding constellation distortion array that will lead the apparatus performing and / or controlling the method according to the first, second and / or third example aspect to satisfy the RF requirement parameter.

[0058] Determining the constellation distortion array may comprise selecting a pre-defined constellation distortion array, for instance from a storage structure such as a table or an array of constellation distortion arrays. For instance, at the apparatus performing and / or controlling the method according to the first, second and / or third example aspect, such a storage structure may be stored (e.g., in a memory]. Additionally or alternatively, within the storage structure, a respective constellation distortion array may be associated with RF requirement parameter values. For instance, respective pre-defined constellation distortion arrays may be configured to conform with respective one or more RF requirement parameters and may thus be retrievable from the storage structure based on one or more RF requirement parameters. The storage structure may additionally or alternatively comprise constellation distortion arrays for different constellations to be distorted, for instance differently sized reference constellations (e.g., different size QAM constellations]. The storage structure may additionally or alternatively comprise constellation distortion arrays for different transmission powers and / or for different resource settings (e.g., frequency range, subcarriers and / ornumerology] to be used by the apparatus performing and / or controlling the method according to the first, second and / or third example aspect The storage structure may thus be queried according to at least one or more of such parameters to select a suitable constellation distortion array from the storage structure. For instance, the selected constellation distortion array may be utilized and / or provided to the network node as is. Additionally or alternatively, the selected constellation distortion array may be further adapted, e.g., according to at least one of the at least one RF requirement parameter.

[0059] Additionally or alternatively, determining the constellation distortion array may comprise generating the constellation distortion array. For instance, the constellation distortion array may be determined (e.g., generated] in a (e.g., joint with the network node] machine learning approach, e.g., wherein the constellation distortion array is optimized towards the apparatus performing and / or controlling the method according to the first, second and / or third example aspect conforming with the RF requirement parameters. Such learning and / or optimizing of a constellation distortion array may for instance be performed from scratch (e.g., starting from a random and / or pre-defined constellation distortion array which may be selected independently of desired RF requirement parameters]. Alternatively, the learning and / or optimizing may be performed based on a selected (e.g., pre-defined] constellation distortion array, e.g., retrieved from the storage structure, e.g., selected based at least one or more of the at least one RF requirement parameter and / or a known characteristic of a power amplifier of the apparatus performing and / or controlling the method according to the first, second or third example aspect.

[0060] According to the first example aspect, the method further comprises providing an indication of the determined constellation distortion array to the network node. For instance, the indication may comprise sufficient information for the network node to construct the constellation distortion array based on the information. For instance, the indication of the constellation distortion array may correspond to the constellation distortion array itself. Additionally or alternatively, the indication of the constellation distortion array may make reference to at least one piece of information (e.g., a known custom constellation] stored at the network node based on which, in addition with the indication of the constellation distortion array, the network node is enabled to derive the constellation distortion array.

[0061] According to the second example aspect, the method comprises providing (e.g., transmitting] a calibration signal to a network node using a reference constellation.

[0062] For instance, the calibration signal may be transmitted at a first transmission power. The transmission power may be a transmission power at the apparatus performing and / or controlling the method according to the first, second and / or third example aspect (e.g., a sending power, e.g., at a given input power]. The first transmission power may for instance be configured to operate a power amplifier of the apparatus performing and / or controlling the method according to the first, second or third example aspect within a transmission chain of the apparatus performing and / or controlling the method according to the first, second or third example aspect used for transmitting the calibration signal in a non-linear range. The transmission power may thus be defined by means of an amplification factor, wherein the amplification factor indicates a ratiobetween a power of an input signal and a power of an output signal of the power amplifier. The first transmission power may for instance be pre-defined (e.g., standard-defined]. The first transmission power may additionally or alternatively be indicated by the network node.

[0063] The calibration signal is modulated according to a reference constellation. A constellation (e.g., reference and / or custom] may be defined by a modulation scheme. Additionally or alternatively, a constellation may be defined as a distribution of constellation points in an in-phase and quadrature, I-Q, plane. A constellation may for instance comprise 2Nconstellation points, e.g., at least or at most or exactly 4, 8, 16, 32, 64, 128, 256, 512, or 1024 constellation points. For instance, a constellation may be symmetric, e.g., axis symmetric about the I and / or Q axis and / or point symmetric about an origin (e.g., point 0,0] of the I-Q-plane. A given constellation point may correspond to a respective binary symbol, for instance a binary symbol comprising N binary bits. For instance, in a constellation with 2N=25=32 constellation points, a given constellation point may represent a given binary symbol expressable as a binary word of N=5-bit length.

[0064] For instance, the reference constellation may correspond to a pre-defined constellation, for instance independent from the apparatus performing and / or controlling the method according to the first, second or third example aspect (e.g., a generic, terminal-device-unspecific constellation]. Additionally or alternatively, the reference constellation may be regular (e.g., comprising evenly spaced constellation points in I-Q-plane], for instance according to a QAM or QPSK constellation. For instance, a respective distance between a given constellation point and a corresponding nearest neighbor constellation point may be identical across all points of the reference constellation.

[0065] The reference constellation used for the calibration signal may be selected by the apparatus performing and / or controlling the method according to the first, second and / or third example aspect. Additionally or alternatively, the reference constellation may be indicated by the network node.

[0066] The calibration signal may have a binary symbol count K, for instance per subcarrier and / or aggregated across subcarriers. For instance, the binary symbol count K may be at least or at most or exactly 256, 512, 1024, 2048, 4096, 2A12, 2A13, or 2A14.

[0067] The calibration signal may have a modulation symbol count L, wherein the modulation symbol count may correspond to a number of subcarriers utilized by the calibration signal. The calibration signal may thus span across multiple subcarriers. For instance, the modulation symbol count L may be at least or at most 2, 3, 4, 5, 6, 7, 8, 16, 32, 64, 128, 256, 512, or 1024.

[0068] The calibration signal may be transmitted according to resource settings (e.g., frequency range, subcarriers and / or numerology (e.g., subcarrier spacing and / or cycling prefix]] to a (e.g., future] data channel, e.g., a Physical Uplink SHared Channel, PUSCH. Additionally or alternatively, the calibration signal may span multiple subcarriers. Such properties of the calibration signal may be entirely or partly chosen by the apparatus performing and / or controlling the method according to the first, second and / or third exampleaspect. Additionally alternatively, such properties of the calibration signal may be entirely or partly indicated by the network node.

[0069] For instance, the number of binary symbols of the calibration signal may be set depending on a number of constellation points in the reference constellation (e.g., and / or the custom constellation]. For instance, the number of binary symbols may be set according to a pre-defined factor applied to the number of constellation points, for instance a factor or at least or at most 2, 3, 4, 5, 8, 16, 32, 64, 128, 256, 512, or 1024. For instance, the number of binary symbols is at least equal to or higher than a number of constellation points in the reference (e.g., and / or custom] constellation. Additionally or alternatively, a number of binary symbols may be set according to (e.g., multiplied by] a number of subcarriers (e.g., modulation symbol count] used in the calibration signal transmission.

[0070] According to the second example aspect, the method further comprises obtaining an indication of a constellation distortion array (e.g., based on the calibration signal] from the network node in response to providing the calibration signal.

[0071] A respective entry of the constellation distortion array indicates a shift in an In-phase / Quadrature, I / Q, plane of a respective reference constellation point of the reference constellation (e.g., used for transmitting the calibration signal] into a corresponding custom constellation point of a custom constellation.

[0072] The constellation distortion array may in case of the second example aspect have been determined by the network node based on the calibration signal, e.g., based on a non-linear characteristic of a power amplifier of the apparatus performing and / or controlling the method according to the second example aspect revealed by the calibration signal transmission.

[0073] According to the third example aspect, the method comprises obtaining (e.g., an indication of] a constellation distortion array. A respective entry of the constellation distortion array indicates a shift in an In-phase / Quadrature, I / Q, plane of a respective reference constellation point of a reference constellation into a corresponding custom constellation point of a custom constellation. For instance, for a reference constellation, the constellation distortion array may indicate respective shifts for all reference constellation points of the reference constellation. The constellation distortion array may be obtained by the apparatus performing and / or controlling the method according to the third aspect determining the constellation distortion array. Additionally or alternatively, the constellation distortion array may be obtained by the apparatus performing and / or controlling the method according to the third aspect receiving the constellation distortion array from a network node.

[0074] According to the third example aspect, the method further comprises modulating uplink information using the reference constellation (e.g., QAM] into at least one reference constellation point of the I / Q plane. The uplink information may for instance comprise a series of bits of which a word of N (e.g., 5] bits is mapped to a single constellation point of 2N(e.g., 25= 32] possible constellation points of the reference constellation (e.g.,32-QAM). The at least one reference constellation point and / or custom constellation point may comprise a same number of constellation points. The processing of the individual constellation points corresponding to different respective pieces of uplink information (e.g., different, consecutive words in a bitstream] may be performed on different subcarriers in a same fashion to span all or a subset of such constellation points.

[0075] According to the third example aspect, the method further comprises determining a shift to be applied to the reference constellation point based on the constellation distortion array. The shift may be selected from the constellation distortion array based on the reference constellation point to shift. For instance, the shifts may be ordered in the constellation distortion array and the reference constellation points may be ordered in a corresponding order (e.g., based on their associated binary value, e.g., in ascending order] so that based on the (e.g. identity of the] respective constellation point, the to-be-applied shift can be selected from the constellation distortion array.

[0076] According to the third example aspect, the method further comprises shifting the at least one reference constellation point according to the determined shift into at least one respective custom constellation point of the custom constellation. For instance, the shift may be represented as a vector in I-Q-space by which the reference constellation point is displaced in order to obtain the custom constellation point.

[0077] According to the third example aspect, the method further comprises performing an uplink transmission (e.g., PUSCH] to a network node, wherein the uplink transmission comprises the at least one custom constellation point. The uplink transmission may comprise a series of custom constellation points respectively obtained by means of the above steps of a mutually different section of a series of bits to be transmitted.

[0078] According to an embodiment of the any of the first, second or third example aspect, the shift is specified in Cartesian coordinates (e.g., (Ai, Aq] for component I and Q] (e.g., two respective coordinates per constellation point] or in Polar coordinates (e.g., (Ar, A<p] for (e.g., change in] radius and angle of a respective constellation point; e.g., with reference constellation point c = (i, q], custom constellation point may be defined as c’ = (i + Ai, q + Aq]; e.g., with reference constellation point c = r ■ el<p, custom constellation point may be defined as c' = (r + Ar) ■ e‘(‘p+A‘p)).

[0079] The constellation distortion array may specify multiple shifts for reference constellation points to arrive at custom constellation points. A respective of such shifts may be defined as a difference vector between the origin (e.g., reference constellation points) and the destination (e.g., custom constellation point). The shift may thus for instance be defined in Cartesian coordinates (e.g., two coordinates, one along the I-axis the other one along the Q-axis in I-Q-plane). Additionally or alternatively, the shift may for instance be defined in polar coordinates (e.g., two coordinates, one indicating an angle about the origin of the I-Q-plane, the other indicating a radius as a distance from the origin of the I-Q-plane). For instance, a shift in Cartesian coordinates of (0.1, 0.1), i.e., +0.1 in the I-axis and +0.1 along the Q-axis may be expressed as ( 0.02, 45°) in polar coordinates. The type of coordinates used may be pre-defined, selected by the apparatus performing and / or controlling the method according to the first, second and / or third example aspect and / or signaled bythe network node. It has been recognized that for certain types of distortions, e.g., twists about the origin of the I-Q-plane, polar coordinates may allow the respective shifts to be expressed by values mutually less different from one another compared to cartesian coordinates.

[0080] The shifts of the constellation distortion array may for instance be expressed and / or encoded (e.g., when indicating the constellation distortion array to another entity] in relative (e.g., difference-] coordinates with respect to either a reference shift (e.g., minimum, maximum shift and / or mean shift across (e.g., all of] the shifts of the constellation distortion array] and / or to one another. Similar values of shifts (e.g., before expressing the shifts relatively] may thus allow for small relative (e.g., difference] values.

[0081] Combining expression of shifts as relative values with expression of shifts using polar coordinates may allow for a particularly efficient representation.

[0082] According to an embodiment of any of the first, second or third example aspect, at least one of

[0083] the reference constellation comprises uniformly distributed reference constellation points;

[0084] the reference constellation corresponds to a square Quadrature Amplitude Modulation, QAM, constellation (e.g., 4-QAM, 16-QAM,..., 2A2m-QAM]; or

[0085] the reference constellation corresponds to another custom constellation.

[0086] A uniform distribution may in particular relate to identical distances between given constellation points and respective nearest neighboring constellation points. A uniform distribution may additionally or alternatively mean that along a path (e.g., a circle] in I-Q-Plane, respective constellation points are equidistantly spaced. For instance, the reference constellation may correspond to a QPSK constellation and / or may conform to this definition. A uniform distribution may additionally or alternatively mean that in the two-dimensional in I-Q-Plane, constellation points are positioned according to a regular grid where any neighboring pair of points is spaced with equal mutual distance. For instance, the reference constellation may correspond to a QAM constellation and / or may conform to this definition.

[0087] The reference constellation may correspond to another custom (e.g., learned] constellation (e.g., different from the custom constellation obtained by applying the constellation distortion array applied to the another custom constellation]. The other custom constellation may for instance have properties such as uneven distribution of constellation points disclosed herein for the custom constellation obtained through application of the constellation distortion array. By selecting the reference constellation as a custom constellation, the shifts indicated by the constellation distortion array may be reduced in number of respective magnitude compared to a regular, pre-defined constellation comprising evenly distributed constellation points such as for instance QAM constellations.

[0088] The reference constellation may be selected by the entity (e.g., network node or terminal device] determining the constellation distortion array and indicated (e.g., by a corresponding identifier of the reference constellation] to the respective other entity (e.g., jointly with the constellation distortion array].According to an embodiment of any of the first, second or third example aspect, at least one of the custom constellation (e.g., and / or the other custom constellation] comprises unevenly distributed custom constellation points (e.g., density of constellation points in I / Q-space not constant across I / Q coordinates]; or

[0089] the custom constellation (e.g., and / or the other custom constellation] corresponds to a learned constellation (e.g., obtained through joint optimization of Block Error Rate, BLER](e.g., obtained through machine learning].

[0090] The custom constellation may be less regular and / or less evenly distributed in I-Q-plane compared to the reference constellation. For instance, constellation points of the custom constellation may be unevenly distributed in the I-Q-plane. This may in particular denote that distances to respective nearest neighboring constellation points vary across constellation points.

[0091] For instance, the custom constellation may be specific to the apparatus performing and / or controlling the method according to the first example aspect, e.g., to (e.g., configured to counteract] a non-linearity of a power amplifier of the apparatus performing and / or controlling the method according to the first, second or third example aspect. Additionally or alternatively, the custom constellation may be configured to be specific to a communication link between the apparatus performing and / or controlling the method according to the first, second or third example aspect and the network node.

[0092] The custom constellation may for instance be a learned constellation. For instance, the custom constellation may be optimized to maximize a communication link property such as for instance spectral efficiency and / or minimize an error rate such as the Block Error Rate, BLER, and / or ensure compliance with the at least one RF requirement parameter by a transmission between the apparatus performing and / or controlling the method according to the first, second or third example aspect and the network node, in particular in uplink, in particular for uplink data transmission in a data channel such as the PUSCH.

[0093] For instance, the custom and / or the reference constellation may be symmetric about the I-axis of the I-Q-plane and / or about the Q-axis in the I-Q-plane and / or about the origin ofthe I-Q-plane.

[0094] According to an embodiment of any of the first, second or third example aspect, the constellation distortion array indicates at least two mutually different shifts for at least two respective different reference constellation points of the reference constellation (e.g., a first for one or more first constellation points and a second for one or more second constellation points].

[0095] One same shift may apply to (e.g., be used to shift] multiple mutually different constellation points of the reference constellation. Yet, the constellation distortion array comprises more than one shift and at least comprises two mutually different shifts that apply to different (e.g., groups of) reference constellation points. Additionally or alternatively, the constellation distortion array may comprise at least two identical shifts.Additionally or alternatively, all shifts comprised by the constellation distortion array may be mutually different. Preferably, the constellation distortion array may comprise (e.g., and / or indicate] as many, half as many or quarter as many shifts as there are constellation points in the reference constellation and / or all such shifts are different from every other shift, respectively.

[0096] According to an embodiment of any of the first, second or third example aspect, the constellation distortion array indicates respective shifts only for a subset of all the reference constellation points of the reference constellation (e.g., in symmetric constellations half or quarter].

[0097] For instance, (e.g., in symmetric constellations symmetric about one (e.g., single] axis of the I-Q-plane] the constellation distortion array may indicate shifts for a half of the I-Q-plane (e.g., for a half defined as I>=0 or a half defined as Q>=0], In this case, the shifts may be applied to reference constellation points of a first half (e.g., I>=0 or Q>=0] as they are indicated in the constellation distortion array while in the corresponding other second half (e.g., I<0 or Q<0], shifts are derived from the shifts indicated for the first half (e.g., by utilizing the same values or by mirroring the shifts about the I- and / or Q-axis],

[0098] Similarly, (e.g., in symmetric constellations symmetric about both axes (i.e., I- and Q-axis] of the I-Q-plane] the constellation distortion array may indicate shifts for a quarter (e.g., a quadrant] of the I-Q-plane and such shifts may be applied to reference constellation points in other quarters, e.g., by mirroring the respective shift about the respective axis crossed when starting from the reference constellation point for which the shift is originally indicated by the constellation distortion array and moving to the reference constellation point to shift.

[0099] According to an embodiment of any of the third example aspect, further comprising providing a calibration signal to the network node using a reference constellation, wherein an indication of the constellation distortion array is received from the network node in response to providing the calibration signal.

[0100] The calibration signal of the third example aspect may be configured according to the above-described calibration signal of the second example aspect.

[0101] According to an embodiment of any of the first, second or third example aspect, the method further comprises

[0102] obtaining an indication of a calibration signal configuration from the network node (e.g., prior to providing the calibration signal]; and

[0103] configuring at least one property of the calibration signal based on the obtained calibration signal configuration.

[0104] The calibration signal configuration may be received from the network node, for instance by at least one of a DCI, a MAC CE and / or an RRC signaling. For instance, the calibration signal configuration may be configuredto trigger a transmission of a calibration signal by the apparatus performing and / or controlling the method according to the first, second and / or third example aspect.

[0105] The calibration signal configuration may indicate at least one property of the calibration signal. For instance, the calibration signal configuration may indicate at least one of a (e.g., reference] constellation (e.g., type of constellation, e.g., regular or irregular; legacy (e.g., QAM] or custom] (e.g., size of constellation, e.g., number of constellation points] to be used when transmitting the calibration signal, a transmission power to be set at the apparatus performing and / or controlling the method according to the first, second and / or third example aspect, a binary symbol count to be transmitted, one or more resource settings (e.g., frequency range, subcarriers and / or numerology (e.g., subcarrier spacing and / or cycling prefix]] (e.g., according to a (e.g., future] data channel, e.g., a Physical Uplink SHared Channel, PUSCH], and / or a number of subcarriers (e.g., modulation symbol count].

[0106] According to an embodiment of any of the first, second or third example aspect, at least one of

[0107] the calibration signal configuration indicates a number of reference constellation points of the reference constellation to be used for the calibration signal (e.g., 4, 8, 16, 32, 64, 128, 256, 256, 512, 1024];

[0108] the calibration signal configuration indicates a type of a (e.g., reference] constellation to be used for the calibration signal (e.g., evenly distributed constellation points, unevenly distributed constellation points, QAM, or learned constellation]; or

[0109] the calibration signal configuration indicates a transmission power to be used by the terminal device (e.g., by a power amplifier of the terminal device] when providing the calibration signal.

[0110] According to an embodiment of any of the first, second or third example aspect, the obtaining the constellation distortion array comprises determining the constellation distortion array.

[0111] According to an embodiment of the third example aspect, the method further comprises obtaining at least one radio frequency, RF, requirement parameter from the network node, wherein the determining the constellation distortion array is based on the obtained at least one RF requirement parameter.

[0112] According to an embodiment of the third example aspect, the obtaining the constellation distortion array comprises receiving the constellation distortion array from the network node.

[0113] The constellation distortion array may have been determined (e.g., selected and / or generated; e.g., as a product of a (e.g., joint] machine learning optimization of a custom constellation] by the network node. The network node may in this case indicate the constellation distortion array to the apparatus performing and / or controlling the method according to the first, second and / or third example aspectAccording to an embodiment of the any of the first, second or third example aspect, the constellation distortion array is determined (e.g., by the network node] (e.g., further, e.g., in addition to the calibration signal] based on (e.g., so that it fulfills] at least one RF requirement parameter.

[0114] The determining the constellation distortion array may for instance be performed based on the RF requirement parameter, in particular in a way that ensures that the apparatus performing and / or controlling the method according to the first, second and / or third example aspect complies with the RF requirement parameter when performing an uplink transmission using a custom constellation obtained by applying the constellation distortion array to the reference constellation. The determining may additionally or alternatively (e.g., in particular if the network node determines the constellation distortion array] be based on a calibration signal provided by the apparatus performing and / or controlling the method according to the first, second and / or third example aspect to the network node.

[0115] According to an embodiment of any of the first, second or third example aspect, the at least one RF requirement parameter comprises a respective requirement, constraint and / or limit (e.g., a threshold] to be fulfilled and / or respected by the terminal device when transmitting a radio signal using the custom constellation.

[0116] For a given RF parameter such as Adjacent Channel Leakage Ratio, ACLR, Peak-to-Average Power Ratio, PAPR, Spectrum Emission Mask, SEM, in-band emissions, Error Vector Magnitude, EVM, and transmit intermodulation, a respective threshold to fulfil by the UE may be set (e.g., by the network node].

[0117] According to an embodiment of any of the first, second or third example aspect, the at least one RF requirement parameter comprises at least one of:

[0118] an out-of-band interference requirement parameter;

[0119] an in-band interference requirement parameter; or

[0120] a spectral mask transmission requirement parameter (e.g., SEM],

[0121] For instance, an out-of-band interference requirement parameter may comprise at least one of Adjacent Channel Leakage Ratio, ACLR, Spectrum Emission Mask, SEM, or transmit intermodulation (e.g., in adjacent channels].

[0122] For instance, an in-band interference requirement parameter may comprise at least one of Peak-to-Average Power Ratio, PAPR, in-band emissions, Error Vector Magnitude, EVM, or transmit intermodulation.

[0123] According to an embodiment of any of the first, second or third example aspect, the method further comprises obtaining a constellation point count of the custom constellation or a dimension of the constellation distortion array (e.g., from the network node].According to an embodiment of the any of the first, second or third example aspect, determining the constellation distortion array comprises at least one of

[0124] selecting the constellation distortion array from (e.g., memory, e.g., from] a table (e.g., three- dimensional table] (e.g., stored in a memory of the terminal device] based on the at least one RF requirement parameter (e.g., by indexing the table by the at least one distortion parameter] (e.g., and further based on a transmission power]; or

[0125] generating the constellation distortion array based on the at least one RF requirement parameter.

[0126] The selecting may be performed based on pre-defined constellation distortion arrays stored in a storage structure (e.g., a table] at the respective apparatus performing the selecting (e.g., the apparatus performing and / or controlling the method according to the first, second and / or third example aspect or the network node], e.g., by indexing the storage structure with the RF requirement array.

[0127] The generating may for instance comprise a (e.g., joint] machine learning and / or optimization, for instance wherein the learning and / or optimization is configured to ensure that the apparatus performing and / or controlling the method according to the first, second and / or third example aspect complies with the RF requirement parameter in upcoming UL transmissions. The learning and / or optimization may be performed based on at least one signal (e.g., the calibration signal] provided by the apparatus performing and / or controlling the method according to the first, second and / or third example aspect to the network node (e.g., in addition to the RF requirement parameter].

[0128] According to an embodiment of the any of the first, second or third example aspect, at least one of

[0129] the table comprises (e.g., is indicative of] at least 2, 3, 4, 5, 6, 7, 8, 16, 32, 64, 128, 256, 512, 1024 or 2048 constellation distortion arrays;

[0130] the table is specific to the terminal device (e.g., to a power amplifier of the terminal device] or to a class of the terminal device; or

[0131] the table is specific to a transmission power.

[0132] The table may for instance comprise constellation distortion arrays that are based on and / or represent previously learned custom constellations. The table may alternatively store custom constellations. Respective stored constellation distortion arrays and / or custom constellations may have previously been obtained through (e.g., joint; e.g., involving both the apparatus performing and / or controlling the method according to the first, second and / or third example aspect and a network node] optimization and / or learning based on respective RF requirement parameters (e.g., learning and / or optimization targeted at the apparatus performing and / or controlling the method according to the first, second and / or third example aspect complying with at least one RF requirement parameter]. Such learned and / or optimized custom constellations and / or constellation distortion arrays may thus have been stored after learning and / or optimization by at least one of the apparatus performing and / or controlling the method according to the first, second and / or third example aspect or the network node in order to be retrievable without performing a (e.g., same] optimization and / or learning again.The terminal device may for instance belong to a class of terminal devices. The class may for instance be defined by a manufacturer of the terminal device and / or a power amplifier of the terminal device (e.g., a given transistor type and / or a given semiconductor (e.g., material, e.g., III-V-semiconductor or Silicon] utilized]. The table may be specific to such a class of terminal devices.

[0133] The table may for instance be variable in size (e.g., in a count of stored custom constellations and / or constellation distortion arrays]. For instance, new constellation distortion arrays and / or custom constellations may be added to the table after a (e.g., joint] learning and / or optimization has been performed. For instance, additionally or alternatively, constellation distortion arrays and / or custom constellations may be removed from the table after a pre-defined period of time such as for instance at least or at most lh, 2h, 5h, 12h, Id, 2d, 5d, lOd, 20d, 30d, or lm. Additionally or alternatively, the table may have a maximum storage capacity, wherein for instance the oldest and / or least utilized stored constellation distortion array and / or custom constellation is removed when a new constellation distortion array and / or custom constellation is added to the table.

[0134] According to an embodiment of any of the first, second or third example aspect, the indication of the constellation distortion array corresponds (e.g., and / or comprises] to the constellation distortion array.

[0135] The constellation distortion array itself may thus be transmitted, in contrast to an indication of the constellation distortion array being transmitted.

[0136] According to an embodiment of the any of the first, second or third example aspect, the method further comprises (e.g., prior obtaining the constellation request] providing a constellation distortion capability indication to the network node, wherein the constellation distortion capability indicates a capability to at least one of

[0137] utilize a constellation distortion array for transforming a constellation (e.g., obtain a custom constellation from a reference constellation];

[0138] determine a constellation distortion array based on at least one RF requirement parameter; or provide a constellation distortion array to the network node.

[0139] Not every apparatus may be able to utilize constellation distortion array in order to obtain and utilize custom constellations. Thus, the apparatus performing and / or controlling the method according to the first, second and / or third example aspect may provide (e.g., transmit] a constellation distortion capability to the network node, e.g., prior to performing steps related to constellation distortion array disclosed herein.

[0140] The constellation distortion capability may for instance be provided to the network node prior to obtaining the RF requirement parameter and / or the constellation request, e.g., prior to providing a calibration signal to the network node and / or e.g., prior to obtaining an indication of a constellation distortion array from the network node. For instance, the apparatus performing and / or controlling the method according to the first,second or third example aspect may provide the constellation distortion capability as a UE Capability Indication (e.g., a Radio Resource Control, RRC, Message], For instance, such a signaling may be triggered by the network node, e.g., via a UECapabilityEnquiry. The apparatus performing and / or controlling the method according to the first, second or third example aspect may respond with a UECapabilitylnformation comprising the constellation distortion capability. Additionally or alternatively, the constellation distortion capability may be indicated by means of an Attach Request and / or Registration Request (e.g., containing UE Network Capability], an Evolved Session Management, ESM, message and / or an RRC message carrying a Non-Access Stratum, NAS, payload.

[0141] The constellation distortion capability may indicate support for utilizing constellation distortion arrays for obtaining custom constellations for uplink transmissions. For instance, the constellation distortion capability may additionally or alternatively indicate a support for participating in an optimization and / or learning process, e.g., involving both the apparatus performing and / or controlling the method according to the first, second or third example aspect and the network node, for obtaining a learned constellation.

[0142] According to an embodiment of the any of the first, second or third example aspect, the method further comprises performing at least one uplink transmission (e.g., PUSCH] to the network node, wherein the at least one uplink transmission is modulated using the custom constellation.

[0143] The apparatus performing and / or controlling the method according to the first, second and / or third example aspect may be configured to transform at least one part of a series of bits to a binary symbol which in turn is converted to a corresponding reference constellation point in I-Q-plane according to a reference constellation and subsequently distorted to a corresponding custom constellation point based on the constellation distortion array. The custom constellation point is part of a modulated radio signal transmitted in uplink from the apparatus performing and / or controlling the method according to the first, second and / or third example aspect to the network node.

[0144] According to an embodiment of the any of the first, second or third example aspect, the modulating comprises applying a normalization factor to the (e.g., shifted custom constellation points of the] custom constellation (e.g., to (e.g., all] constellation points of the custom constellation], wherein the normalization factor is based on (e.g., adjusted by] the constellation distortion array (e.g., in order to achieve and / or thereby achieving compliance with a pre-defined power constraint, e.g., unity power in average across all constellation points].

[0145] Normalization in the context of constellations may for instance refer to adjusting a length (e.g., representing a square root of the power] of a respective constellation point vector (e.g., from origin of I-Q-plane to the respective constellation point] by a (e.g., single] normalization factor (e.g., same for all constellation points]. Normalization may in particular be performed to obtain an average power across the constellation points of the constellation (e.g., of the transmitted binary symbols] that equates to a pre-defined value, such as for instance unity power of one. This ensures that the constellation does not influence an average transmissionpower (e.g. in case of equiprobable binary symbols] and / or for instance, performance metrics such as error vector magnitude and / or signal-to-noise ratio may be compared across constellations easily.

[0146] It has been recognized that custom constellations and in particular distorted constellations that are a result of applying the constellation distortion array to a reference constellation may cause require an (e.g., additional] normalization using a normalization factor that may be set based on the constellation distortion array. For instance, the normalization factor may additionally be based on the reference constellation. For instance, the normalization factor may be obtained by computing a sum of the (e.g., squared] lengths of custom constellation points obtained by applying the constellation distortion array and dividing the sum by the total number of constellation points in the constellation (assuming equal probabilities for all the constellation points]. Each custom constellation point may subsequently be scaled (e.g., normalized] by a factor of one over the divided sum.

[0147] According to an embodiment, the constellation distortion array may itself be normalized, e.g., for a respective reference constellation. For instance, respective influences of the shifts on respective powers of custom constellation points as caused by the constellation distortion array may be selected to even one another out (e.g., across all custom constellation points] so that on average across custom constellation points, no change in average power is caused by the constellation distortion array.

[0148] According to a fourth example aspect, a method is disclosed, for instance performed by a network node (e.g., gNB], comprising:

[0149] providing at least one radio frequency, RF, requirement parameter to a terminal device; providing a constellation request for a custom constellation to the terminal device; and

[0150] obtaining an indication of a constellation distortion array (e.g., a matrix, e.g. of dimension 2xN] from the terminal device, wherein a respective entry of the constellation distortion array indicates a shift in an In-phase / Quadrature, I / Q, plane of a respective reference constellation point of a reference constellation into a corresponding custom constellation point of the custom constellation.

[0151] According to a fifth example aspect, a method is disclosed, for instance performed by a network node (e.g., gNB], comprising:

[0152] obtaining a calibration signal from a terminal device based on a reference constellation; determining a constellation distortion array at least partially based on the obtained calibration signal, wherein a respective entry of the constellation distortion array indicates a shift in an In- phase / Quadrature, I / Q, plane of a respective reference constellation point of the reference constellation into a corresponding custom constellation point of a custom constellation; and providing an indication of the determined constellation distortion array to the terminal device.

[0153] The method according to the fourth or fifth example aspect may for instance be performed and / or controlled by an apparatus, for instance a server. Alternatively, the method may be performed and / or controlled by more than one apparatus, for instance a server cloud comprising at least two servers. Alternatively, themethod may for instance be performed and / or controlled by an electronic device, e.g. a node in a communication system. For instance, the method may be performed and / or controlled by using at least one processor of the electronic device.

[0154] According to a further example aspect, a computer program is disclosed, the computer program when executed by a processor causing an apparatus, for instance a server, a network node or a terminal device, e.g., a UE, to perform and / or control the actions of the method according to the fourth or fifth example aspect.

[0155] The computer program may be stored on computer-readable storage medium, in particular a tangible and / or non-transitory medium. The computer readable storage medium could for example be a disk or a memory or the like. The computer program could be stored in the computer readable storage medium in the form of instructions encoding the computer-readable storage medium. The computer readable storage medium may be intended for taking part in the operation of a device, like an internal or external memory, for instance a Read-(e.g., Only] Memory [ROM] or hard disk of a computer, or be intended for distribution of the program, like an optical disc.

[0156] According to a further aspect, a computer program product is disclosed comprising the computer program.

[0157] According to a further example aspect, an apparatus is disclosed, configured to perform and / or control or comprising respective means for performing and / or controlling the method according to the fourth or fifth example aspect. The means may for instance comprise a power amplifier, for instance in a transmission chain for uplink transmissions to a network node.

[0158] The means of the apparatus can be implemented in hardware and / or software. They may comprise for instance at least one processor for executing computer program code for performing the required functions, at least one memory storing the program code, or both. Alternatively, they could comprise for instance circuitry that is designed to implement the required functions, for instance implemented in a chipset or a chip, like an integrated circuit. In general, the means may comprise for instance one or more processing means or processors.

[0159] The above-disclosed apparatus according to any aspect may be a module or a component for a device, for example a chip. Alternatively, the disclosed apparatus according to any aspect may be a device, for instance a server or server cloud. The disclosed apparatus according to any aspect may comprise (e.g., only] the disclosed components, for instance means, processor, memory, or may further comprise one or more additional components.

[0160] According to a further example aspect, an apparatus, e.g., a network node, is disclosed, comprising respective means for performing the actions of the method of any of the fourth or fifth example aspect, for instance according to at least one or more embodiments of the fourth or fifth example aspect.According to a further example aspect, an apparatus, e.g., a network node, is disclosed, comprising at least one processor; and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to perform and / or control the actions of the method according any of embodiments of any of the fourth or fifth example aspect, for instance according to at least one or more embodiments of the fourth or fifth example aspect.

[0161] Features disclosed with respect to the first, second or third example aspect are disclosed for the fourth and fifth example aspect as well, where applicable.

[0162] An obtaining of a quantity (e.g., the indication of the constellation distortion array] by the apparatus performing and / or controlling the method according to the fourth or fifth example aspect, e.g., from the terminal device, may comprise and / or correspond to receiving the quantity or an according indication of the quantity from the terminal device. The receiving may be performed by means of at least one of a Uplink Control Information, UCI, a Medium Access Control Control Element, MAC CE, a Radio Resource Control, RRC, signaling and / or combinations thereof.

[0163] A providing of a quantity (e.g., the RF requirement parameter, the constellation request, the indication of a constellation distortion array and / or the constellation distortion array] by the apparatus performing and / or controlling the method according to the fourth or fifth example aspect to the terminal device may comprise and / or correspond to transmitting the quantity or an according indication of the quantity to the terminal device. The transmitting may be performed by means of at least one of an Uplink Control Information, UCI, a Medium Access Control Control Element, MAC CE, a Radio Resource Control, RRC, signaling and / or combinations thereof.

[0164] According to an embodiment of the fourth or fifth example aspect, the shift is specified in Cartesian coordinates (e.g., (Ai, Aq] for component I and Q] (e.g., two respective coordinates per constellation point] or in Polar coordinates (e.g., (Ar, A<p] for (e.g., change in] radius and angle of a respective constellation point; e.g., with reference constellation point c = (i, q], custom constellation point may be defined as c’ = (i + Ai, q + Aq]; e.g., with reference constellation point c = r ■ el<p, custom constellation point may be defined as c' = (r + Ar) ■ e‘(‘p+A‘p)).

[0165] According to an embodiment of the fourth or fifth example aspect, at least one of

[0166] the reference constellation comprises uniformly distributed reference constellation points;

[0167] the reference constellation corresponds to a square Quadrature Amplitude Modulation, QAM, constellation (e.g., 4-QAM, 16-QAM,..., 22m-QAM); or

[0168] the reference constellation corresponds to another custom constellation.

[0169] According to an embodiment of the fourth or fifth example aspect, at least one ofthe custom constellation (e.g., and / or the other custom constellation] comprises unevenly distributed custom constellation points (e.g., density of constellation points in I / Q-space not constant across I / Q coordinates]; or

[0170] the custom constellation (e.g., and / or the other custom constellation] corresponds to a learned constellation (e.g., obtained through joint optimization of Block Error Rate, BLER](e.g., obtained through machine learning].

[0171] According to an embodiment of the fourth or fifth example aspect, the constellation distortion array indicates at least two mutually different shifts for at least two respective different reference constellation points of the reference constellation (e.g., a first for a first one or more constellation points and a second for a second one or more constellation points].

[0172] According to an embodiment of the fourth or fifth example aspect, the constellation distortion array indicates respective shifts only for a subset of all the reference constellation points of the reference constellation (e.g., in symmetric constellations half or quarter].

[0173] According to an embodiment of the fourth or fifth example aspect, the method further comprises providing an indication of a calibration signal configuration to the terminal device (e.g., prior to providing the calibration signal], wherein the calibration signal configuration indicates at least one property of the calibration signal.

[0174] According to an embodiment of the fourth or fifth example aspect, at least one of

[0175] the calibration signal configuration indicates a number of reference constellation points of the reference constellation to be used for the calibration signal (e.g., 4, 8, 16, 32, 64, 128, 256, 256, 512, 1024];

[0176] the calibration signal configuration indicates a type of a (e.g., reference] constellation to be used for the calibration signal (e.g., evenly distributed constellation points, unevenly distributed constellation points, QAM, or learned constellation]; or

[0177] the calibration signal configuration indicates a transmission power to be used by the terminal device (e.g., by a power amplifier of the terminal device] when providing the calibration signal.

[0178] According to an embodiment of the fourth or fifth example aspect, wherein the determining the constellation distortion array is further (e.g., in addition to the calibration signal] based on at least one radio frequency, RF, requirement.

[0179] According to an embodiment of the fourth or fifth example aspect, the at least one RF requirement parameter comprises a respective requirement (e.g., a threshold] to be fulfilled by the terminal device when transmitting a radio signal using the custom constellation.1

[0180] According to an embodiment of the fourth or fifth example aspect, the at least one RF requirement parameter comprises at least one of:

[0181] an out-of-band interference requirement parameter;

[0182] an in-band interference requirement parameter; or

[0183] a spectral mask transmission requirement parameter.

[0184] According to an embodiment of the fourth or fifth example aspect, the method further comprises providing a constellation point count of the custom constellation or a dimension of the constellation distortion array to the terminal device.

[0185] According to an embodiment of the fourth or fifth example aspect, the indication of the constellation distortion array corresponds (e.g., and / or comprises] to the constellation distortion array.

[0186] According to an embodiment of the fourth or fifth example aspect, the method further comprises (e.g., prior obtaining the constellation request] obtaining a constellation distortion capability indication from the terminal device, wherein the constellation distortion capability indicates a capability to at least one of utilize a constellation distortion array for transforming a constellation;

[0187] determine a constellation distortion array based on at least one RF requirement parameter; or provide a constellation distortion array to the network node.

[0188] According to an embodiment of the fourth or fifth example aspect, the determining the constellation distortion array comprises at least one of (e.g., based on the at least one distortion parameter] generating (e.g., through a machine learning model and / or an optimization algorithm] a constellation distortion array; or

[0189] selecting (e.g., a basis constellation distortion array for] a constellation distortion array (e.g., from a table].

[0190] The apparatus performing and / or controlling the method according to the fourth or fifth example aspect may generate a constellation distortion array by performing and / or controlling a (e.g., joint; e.g., in cooperation with the terminal device, e.g., the apparatus performing and / or controlling the method according to the first, second and / or third example aspect] an optimization and / or learning procedure. For instance, the generating may be based on a calibration signal obtained from the terminal device. The learning and / or optimizing may ensure that the terminal device complies with an RF requirement parameter in a future uplink transmission.

[0191] Additionally or alternatively, the constellation distortion array may be selected by the apparatus performing and / or controlling the method according to the fourth or fifth example aspect For instance, the apparatus performing and / or controlling the method according to the fourth or fifth example aspect may store a storage structure as disclosed above for the apparatus performing and / or controlling the method according to the first, second and / or third example aspect such as a table. For instance, the apparatus performing and / or controlling the method according to the fourth or fifth example aspect may store custom constellationsand / or constellation distortion arrays for multiple mutually different terminal devices, for instance with which it has previously performed a (e.g., joint] learning and / or optimization of a respective custom constellation and / or constellation distortion array. After such learning and / or optimization, the apparatus performing and / or controlling the method according to the fourth or fifth example aspect may store the custom constellation and / or constellation distortion array for future usage and select it when required. For instance, the apparatus performing and / or controlling the method according to the fourth or fifth example aspect may receive a calibration signal from the apparatus performing and / or controlling the method according to the first, second and / or third example aspect and determine that the apparatus performing and / or controlling the method according to the first, second and / or third example aspect (e.g., a power amplifier of the apparatus performing and / or controlling the method according to the first, second and / or third example aspect] behaves similarly to another terminal device for which a custom constellation and / or constellation distortion array has been learned and / or optimized (e.g., according to a RF requirement parameter currently desired] and that such custom constellation and / or constellation distortion array is still available in storage. The apparatus performing and / or controlling the method according to the fourth or fifth example aspect may then select the respective custom constellation and / or constellation distortion array and provide it to the apparatus performing and / or controlling the method according to the first, second and / or third example aspect

[0192] According to an embodiment of the fourth or fifth example aspect, selecting the constellation distortion array comprises selecting the constellation distortion array from (e.g., memory, e.g., from] a table (e.g., three- / four- / five-dimensional table] (e.g., stored in a memory of the network node] using the at least one RF requirement parameter (e.g., by indexing the table by the at least one distortion parameter] (e.g., and further using a transmission power].

[0193] According to an embodiment of the fourth or fifth example aspect, at least one of

[0194] the table comprises (e.g., is indicative of] at least 2, 3, 4, 5, 6, 7, 8, 16, 32, 64, 128, 256, 512, 1024 or 2048 constellation distortion matrices]; or

[0195] the table is specific to a transmission power.

[0196] According to an embodiment of the fourth or fifth example aspect, the method further comprises adjusting a receiver of the network node in accordance with the determined constellation distortion array (e.g., in order to be able to receive uplink communication from the terminal device using the constellation distortion array],

[0197] A receiver, e.g., a demapper, may be adjusted to successfully establish correspondence between received constellation points and their bit values.

[0198] According to an embodiment of the fourth or fifth example aspect, the method further comprises receiving at least one uplink transmission (e.g., PUSCH] from the terminal device and demodulating the at least one uplink transmission using the custom constellation (e.g., and / or using the constellation distortion array].The demodulating may in particular comprise demapping at least one custom constellation point to a corresponding series of bits based on the constellation distortion array.

[0199] According to a further example aspect, a system is disclosed comprising at least one first apparatus according to any of the first, second or third example aspect and at least one second apparatus according to any of the fourth or fifth example aspect.

[0200] The features and example embodiments of the invention described above may equally pertain to the different aspects according to the present invention. In particular, disclosure presented with respect to the first or second aspect is also disclosed for the third and fourth aspect, where applicable.

[0201] It is to be understood that the presentation of the invention in this section is merely by way of examples and non-limiting.

[0202] Other features of the invention will become apparent from the following detailed description considered in conjunction with the accompanying drawings. It is to be understood, however, that the drawings are designed solely for purposes of illustration and not as a definition of the limits of the invention, for which reference should be made to the appended claims. It should be further understood that the drawings are not drawn to scale and that they are merely intended to conceptually illustrate the structures and procedures described herein.

[0203] BRIEF DESCRIPTION OF THE DRAWINGS

[0204] In the figures show:

[0205] Fig. 1 a schematic visualization of a communication system according to embodiments of all example aspects;

[0206] Fig. 2 a schematic visualization of regular and custom constellations in I-Q-plane and time and frequency according to embodiments of all example aspects;

[0207] Fig. 3a, b a schematic visualization of a regular and of a custom constellation according to embodiments of all example aspects;

[0208] Fig. 4 a signaling chart showing a method according to an example embodiment of all example aspects;

[0209] Fig. 5 a signaling chart showing a method according to an example embodiment of all example aspects;

[0210] Fig. 6a, b examples of shifts applied to constellation points according to example embodiments of all example aspects;

[0211] Fig. 7 an example illustration of a constellation distortion array according to example embodiments of all example aspects;

[0212] Fig. 8a, b examples of shifts applied to constellation points of a QAM reference constellation according to example embodiments of all example aspects;Fig. 9 a flowchart showing an example embodiment of a method according to the first example aspect;

[0213] Fig. 10 a flowchart showing an example embodiment of a method according to the second example aspect;

[0214] Fig. 11 a flowchart showing an example embodiment of a method according to the third example aspect;

[0215] Fig. 12 a flowchart showing an example embodiment of a method according to the fourth example aspect;

[0216] Fig. 13 a flowchart showing an example embodiment of a method according to the fifth example aspect;

[0217] Fig. 14 a schematic block diagram of an apparatus configured to perform the method according to the first example aspect;

[0218] Fig. 15 a schematic block diagram of an apparatus configured to perform the method according to the second example aspect;

[0219] Fig. 16 a schematic block diagram of an apparatus configured to perform the method according to the third example aspect;

[0220] Fig. 17 a schematic block diagram of an apparatus configured to perform the method according to the fourth example aspect;

[0221] Fig. 18 a schematic block diagram of an apparatus configured to perform the method according to the fifth example aspect; and

[0222] Fig. 19 a schematic illustration of examples of tangible and non-tangible storage media according to all example aspects.

[0223] Fig. 1 demonstrates a simplified communication system, in which methods according to all aspects may be performed. The system comprises a first terminal device 100, e.g., an apparatus performing and / or controlling the method according to the first, second or third example aspect, and a network node 200, e.g., an apparatus performing and / or controlling the method according to the fourth or fifth example aspect.

[0224] The terminal device 100 and the network node 200 may be configured to agree on a custom constellation by signaling RF requirement parameter (e.g., an in-band and / or out-of-band interference constraint] and / or indications of constellation distortion arrays. Either the terminal device 100 or the network node 200 may be configured to determine a constellation distortion array. The constellation distortion array may for instance be retrieved (e.g., selected] according to the RF requirement parameter and / or may be determined in a joint learning and / or optimization procedure involving both the terminal device 100 and the network node 200. For instance, the terminal device 100 may be configured to provide a calibration signal to the network node 200. The network node 200 may apply a learning and / or optimization of a constellation based on the received calibration signal. For instance, the learning and / or optimization may be performed based on a previously learned and / or optimized custom constellation, e.g., provided by the terminal device 100 to the network node 200 and / or stored at the network node 200.The network node 200 may request the terminal devi8ce 100 to transmit a known uplink calibration signal block using a predetermined modulation scheme (e.g., reference modulation, QAM, and / or a known custom constellation]. By analyzing the received reference signal, the network node 200 may evaluate distortions (e.g., of the constellation, e.g., with respect to the assumed constellation]. Such distortions may in particular be introduced by the terminal device’s 100 power amplifier and / or other impairments, e.g., the radio conditions. The network node 200 may evaluate whether to adjust the modulation format (e.g., the constellation] and / or apply corrections (e.g., of the constellation] to maintain optimal uplink performance.

[0225] A custom constellation may for instance correspond to a constellation for which constellation points are determined by machine learning. In this case, the custom constellation may be termed a learned constellation. Additionally or alternatively, the custom constellation may be a constellation that is not a QAM constellation but have been optimized (e.g., according to a given cost function] through optimization methods.

[0226] The present disclosure therefore relates to methods and / or apparatus for coping with power amplifier (PA] nonlinearity that has a terminal device 100 (e.g., UE] component and a network node 200 (e.g., gNB] component. Specifically, the proposed framework may comprise at least some of the following features. At least one I-Q-transformation and / or distortion to the constellation may be derived and deployed, for instance in the form of a constellation distortion array (e.g., constellation distortion vector or matrix] and / or as a function. The transformation and / or distortion may be targeted towards minimizing at least one RF requirement parameter such as SEM and / or ACLR, for instance caused by the transmitter’s (e.g., terminal device’s] PA non-linearity, e.g., caused by memory effects. The derivation may be done using tools from machine learning and may require training over the air interface for tuning the constellation. It has been recognized that a byproduct of the derivation of a constellation optimizing SEM and / or ACLR may be an additional in-band distortion, i.e. in addition to the in-band distortions introduced by the PA response.

[0227] A (e.g., further] step of deriving a constellation distortion array and / or function under given SEM and ACLR constraints, e.g., and for a UE PA response may for instance be performed by the UE 100 and communicated to the gNB 200. In this case, the SEM and ACLR limits may (e.g., need to] be enforced by the gNB 200. The UE 100 may communicate to the gNB the resulting constellation distortion vector (CDV], each characterized by a target SEM and ACLR

[0228] According to an example embodiment, the terminal device 100 (e.g., and / or the network node 200] may hold available a table of constellation distortion arrays which may for instance correspond to constellation distortion vectors, CDVs. When a CDV is disclosed, the corresponding features is also disclosed for a constellation distortion array, where applicable, A given CDV may be associated with RF requirement parameters (e.g., constraints] such as for instance SEM and / or ACLR. An example is shown in table 1 below.

[0229] Index CDV SEM ACLR

[0230] 1 [a, b, c, ...] SEM-limit [1] ACLR-limit[l]

[0231]

[0232] Table 1: CDV tableCDV may for instance correspond to and / or comprise a vector that is superimposed to a reference constellation (e.g., a QAM constellation] to obtain a custom (e.g., custom QAM] constellation. For instance, to a respective reference (e.g., QAM] constellation point k a distortion and / or shift, e.g., corresponding to k-th entry in the CDV to yield a new custom constellation point.

[0233] Additionally or alternatively to a CDV, a constellation distortion function, CDF, f may be a function that is applied to a respective (e.g., each] QAM point k, to transform it to another point in a custom constellation. For instance, the gNB 200 itself may derive the CDV and / or CDF, e.g., by triggering the UE to transmit training signals (e.g., the calibration signal].

[0234] Deploying the constellation distortion array (e.g., a CDV] may require a coordination between the UE 100 and the gNB 200. An Al enhanced receiver (e.g., at the network node 200] may compensate for a (e.g., total] in-band distortion caused by the PA and the custom constellation (e.g., if any]. The receiver (e.g., at the gNB 200] may be trained after the CDV was obtained, either using the custom constellation and / or offline generated data. The gNB 200 may to this end approximate and / or model a PA response. Alternatively, the receiver may be trained by triggering the UE 100 to transmit a known signal (e.g., calibration signal] with said constellation.

[0235] Fig. 2 shows an effect of (e.g., reduction by] a custom constellation on ACLR and / or PAPR. In the above three plots, the left one shows a regular QAM constellation. In the middle one, a waveform in time of a transmitted signal shows three peaks that for instance depart significantly from an average signal strength. A PAPR may in this case be deemed to be elevated. The right plot shows the corresponding (e.g., power] spectrum in which considerable power is present outside the channel, thereby causing an increase ACLR rating. The lower row illustrates how the custom constellation with irregularly positioned constellation points allows for both PAPR of the OFDM waveform in time and ACLR in the frequency spectrum to be reduced. Such reductio may for instance be caused and / or achieved by minimizing the amount and / or number of peaks in the signal by shaping the custom constellation to achieve this reduction. This constellation shaping can be achieved as an implementation-specific solution for instance. Yet, the way of sharing such constellations between gNB and UE is facilitated particularly as detailed in the present disclosure.

[0236] Fig. 3a shows a reference constellation wherein the constellation points are positioned regularly, e.g., a QAM constellation. Further, Fig. 3b shows a custom constellation, for instance obtained by applying a constellation distortion array (CDA] to the reference constellation of Fig. 3a.

[0237] Fig. 4 shows a signaling diagram of a UE 100 (e.g., according to the first or third example aspect] transferring a constellation distortion array (CDA], e.g., in accordance RF requirement parameters (e.g., leakage constraints] set (e.g., requested] by a gNB 200. In this configuration, the UE 100 may derive the CDA locally, e.g., based on its PA characteristics and the constraints communicated by the gNB 200 (e.g., according to the fourth example aspect].In an (e.g., optional] step S101, the UE 100 may indicate a capability to utilize CDA in general and / or to in particular obtain a CDA, in particular based on RF requirement parameters such as for instance SEM and / or ACLR constraints.

[0238] In a step S102, the gNB 100 sends a request to the UE 100 for a CDA. The request may for instance specify a constellation size and / or RF requirement parameters such as for instance SEM and / or ACLR limits.

[0239] The UE 100 may in step S103 determine the appropriate CDA. For instance, the CDA may be retrieved from a (e.g., pre-computed] table (e.g., stored locally], A respective CDA may be indexed by corresponding target SEM and / or ACLR constraints.

[0240] In step S104, the UE 100 transfers the selected CDA to the gNB 200. The transmitted signal may be ensured to adhere to the specified requirements on, e.g., SEM and / or ACLR.

[0241] The gNB 200 and UE 100 subsequently use the custom constellation in step S105, wherein the custom constellation may be represented as the reference constellation (e.g., standard QAM constellation] with an additional distortion as indicated by the CDA applied to it

[0242] The UE 100 may thus be capable to (e.g., independently] compute and deploy CDAs. This may reduce computational burden on the gNB 200 while ensuring compliance with spectral constraints.

[0243] Figure 5 shows an embodiment wherein a gNB 200 (e.g., according to the fifth example aspect] trains and transfers a CDA toa UE 100 (e.g., according to the second and / or third example aspect].

[0244] In this configuration, the gNB 200 may dynamically train the CDV, e.g., using (e.g., real-time] calibration signals transmitted by the UE 100.

[0245] In step S201, the UE 100 may indicate a capability to support training of a custom constellation (e.g., under SEM and / or ACLR constraints] and / or a capability to receive custom constellations by means of a CDA.

[0246] The gNB 200 requests, in step S202, the UE 100 to transmit a known calibration signal, such as a standard QAM constellation of a specified size (e.g., at a particular transmission power].

[0247] In step S203, the UE 100 transmits the calibration signal, e.g., in accordance with the characteristics requested by the gNB 200 in step S203.

[0248] In steps S204, S205, S206, S207, the gNB 200 uses the received calibration signal to model a PA response of the UE 100 and may further derive a CDA that minimizes SEM and / or ACLR violations. The gNB 200 may further fine-tune the CDA to ensure compatibility with the receiver’s capabilities and the operating environment.The derived CDA is then transferred to the UE in step S208. The UE 100 may apply the CDA to its transmission signal.

[0249] In step S209, the UE 100 and the gNB 200 utilize the custom constellation (e.g., as in the reference constellation QAM + the constellation distortion array CDA] for uplink transmission.

[0250] The approach shown in Fig. 5 may be particularly useful in dynamic and / or non-static environments wherein a pre-computed CDAs may not be sufficient. By training the CDA (e.g., in real time], the system may be able to adapt to changes in PA behavior caused by temperature, aging, or other environmental factors.

[0251] Fig. 6a, b illustrate shifts indicated by the constellation distortion array. In Fig. 6a, the shift v is indicated in Cartesian coordinates, e.g., as (Ai, Aq], A constellation point c = (i, q] may be shifted according to the shift v to a new (e.g., custom] constellation points c’ = (i’, q’] = (i + Ai, q + Aq], Fig. 6b shows the same shift, yet indicated in polar coordinates. For instance, a reference constellation point c = r ■ eltpmay be shifted with a shift v = (Ar, A<p] to custom constellation point c' = r' ■ eltp= (r + Ar) ■ el(<p+A<p).

[0252] Fig. 7 may be used to illustrated symmetries of constellations and corresponding applications of constellation distortion arrays. For instance if a constellation to be distorted and / or to be obtained by distortion is symmetric about the I- and the Q-axis, a CDA may (e.g., only) indicate shifts for a quarter of the constellation points, e.g., for the constellation points in the first quarter I >=0, Q>=0 (CDA). The respective shifts may then be re-used for the remaining three quarters, e.g., while changing the sign of at least one or both of the coordinates of the respective shift. For instance, when the shifts are indicated, respectively, as v=(Ai, Aq) in the quarter I>=0, Q>=0, termed QDA, for the quarter CDA’ (I<0, Q>=0), the i-component of the respective shift may be inverted to v’=(-Ai, Aq). Similarly, for CDA’”, the q-component may be inverted to v’”=(Ai, -Aq). For CDA” (I <0, Q<0), both coordinates may be inverted to v”=(-Ai, -Aq).

[0253] Fig. 8a, b demonstrates how an application of a constellation distortion array may perturb a normalization of a constellation. For instance, when considering a single constellation point Ci, the shifted constellation point Ci’ may be larger (e.g., have a higher length and / or norm) and thus transport more power. The custom constellation obtained as a result of obtaining may thus be required to normalized again by the square root over the sum of powers of the custom constellation points. Additionally or alternatively, the respective influences of the shifts on respective powers of constellation points may be selected to even one another out so that no re-normalization is required. The principle is illustrated for a 4-QAM in Fig. 8a and for a 16-QAM in Fig. 8b.

[0254] Fig. 9 shows a flowchart of an example embodiment according to the first example aspect, e.g., performed and / or controlled by a terminal device 100.In a step M101, the shown method comprises obtaining at least one radio frequency, RF, requirement parameter from a network node (e.g., performing and / or controlling a method according to the fourth or fifth aspect].

[0255] In a step M102, the shown method comprises obtaining a constellation request for a custom constellation from the network node.

[0256] In a step M103, the shown method comprises determining a constellation distortion array (e.g., a matrix, e.g. of dimension 2xN] based on the at least one RF requirement parameter, wherein a respective entry of the constellation distortion array indicates a shift in an In-phase / Quadrature, I / Q, plane of a respective reference constellation point of a reference constellation into a corresponding custom constellation point of the custom constellation.

[0257] In a step M104, the shown method comprises providing an indication of the determined constellation distortion array to the network node.

[0258] Fig. 10 shows a flowchart of an example embodiment according to the second example aspect, e.g., performed and / or controlled by a terminal device 100.

[0259] In a step M201, the shown method comprises providing (e.g., transmitting] a calibration signal to a network node (e.g., performing and / or controlling a method according to the fourth or fifth aspect] using a reference constellation.

[0260] In a step M202, the shown method comprises obtaining an indication of a constellation distortion array (e.g., based on the calibration signal] from the network node in response to providing the calibration signal, wherein a respective entry of the constellation distortion array indicates a shift in an In-phase / Quadrature, I / Q, plane of a respective reference constellation point of the reference constellation into a corresponding custom constellation point of a custom constellation.

[0261] Fig. 11 shows a flowchart of an example embodiment according to the third example aspect, e.g., performed and / or controlled by a terminal device 100.

[0262] In a step M301, the shown method comprises obtaining (e.g., an indication of] a constellation distortion array, wherein a respective entry of the constellation distortion array indicates a shift in an In-phase / Quadrature, I / Q, plane of a respective reference constellation point of a reference constellation into a corresponding custom constellation point of a custom constellation. The CDA may for instance be obtained from memory or from a network node, for instance a network node performing and / or controlling a method according to the fifth aspect.In a step M302, the shown method comprises modulating uplink information using the reference constellation (e.g., QAM] into at least one reference constellation point of the I / Q plane.

[0263] In a step M303, the shown method comprises determining a shift to be applied to the reference constellation point based on the constellation distortion array.

[0264] In a step M304, the shown method comprises shifting the at least one reference constellation point according to the determined shift into at least one respective custom constellation point of the custom constellation.

[0265] In a step M305, the shown method comprises performing an uplink transmission (e.g., PUSCH] to a network node, wherein the uplink transmission comprises the at least one custom constellation point.

[0266] Fig. 12 shows a flowchart of an example embodiment according to the fourth example aspect, e.g., performed and / or controlled by a network node 200.

[0267] In a step M401, the shown method comprises providing at least one radio frequency, RF, requirement parameter to a terminal device (e.g., performing and / or controlling a method according to the first, second or third aspect].

[0268] In a step M402, the shown method comprises a constellation request for a custom constellation to the terminal device.

[0269] In a step M403, the shown method comprises obtaining an indication of a constellation distortion array (e.g., a matrix, e.g. of dimension 2xN] from the terminal device, wherein a respective entry of the constellation distortion array indicates a shift in an In-phase / Quadrature, I / Q, plane of a respective reference constellation point of a reference constellation into a corresponding custom constellation point of the custom constellation.

[0270] Fig. 12 shows a flowchart of an example embodiment according to the fifth example aspect, e.g., performed and / or controlled by a network node 200.

[0271] In a step M501, the shown method comprises obtaining a calibration signal from a terminal device based on a reference constellation.

[0272] In a step M502, the shown method comprises determining a constellation distortion array at least partially based on the obtained calibration signal, wherein a respective entry of the constellation distortion array indicates a shift in an In-phase / Quadrature, I / Q, plane of a respective reference constellation point of the reference constellation into a corresponding custom constellation point of a custom constellation.

[0273] In a step M503, the shown method comprises providing an indication of the determined constellation distortion array to the terminal device.Fig. 14 shows an example block diagram of an apparatus 100 (e.g., performing and / or controlling a method according to the first example aspect, e.g., a terminal device]. The apparatus 100 may perform a method according to the first example aspect. The apparatus may comprise a user interface A160, a program memory A110, amain memory A120, and a data memory A140. Further, it comprises a processor A130. The apparatus 100 may further comprise the functional units RF requirement parameter obtainer A131 (e.g., corresponding to method step M101 of Fig. 9], constellation request obtainer A132 (e.g., corresponding to method step M102 of Fig. 9], constellation distortion array obtainer (e.g., corresponding to method step M103 of Fig. 9], constellation distortion array indication provider A134 (e.g., corresponding to method step M104 of Fig. 9], A functional unit may for instance correspond to a code block within a memory A110, A120, A140. The functional units A131, A132 and A134 may for instance be connected to and / or control the communication interface Al 50. The communication interface Al 50 may for instance comprise a power amplifier.

[0274] Fig. 15 shows an example block diagram of an apparatus 101 (e.g., performing and / or controlling a method according to the second example aspect, e.g., a terminal device]. The apparatus 101 may perform a method according to the second example aspect. The apparatus may comprise a user interface A260, a program memory A210, amain memory A220, and a data memory A240. Further, it comprises a processor A230. The apparatus 101 may further comprise the functional units calibration signal provider A231 (e.g., corresponding to method step M201 of Fig. 10], constellation distortion array indication obtainer A232 (e.g., corresponding to method step M202 of Fig. 10], A functional unit may for instance correspond to a code block within a memory A210, A220, A240. The functional units A231, A232 may for instance be connected to and / or control the communication interface A250. The communication interface A250 may for instance comprise a power amplifier.

[0275] Fig. 16 shows an example block diagram of an apparatus 102 (e.g., performing and / or controlling a method according to the third example aspect, e.g., a terminal device]. The apparatus 102 may perform a method according to the third example aspect. The apparatus may comprise a user interface A360, a program memory A310, amain memory A320, and a data memory A340. Further, it comprises a processor A330. The apparatus 102 may further comprise the functional units constellation distortion array obtainer A331 (e.g., corresponding to method step M301 of Fig. 11], uplink information modulator A332 (e.g., corresponding to method step M302 of Fig. 11], shift determiner A333 (e.g., corresponding to method step M303 of Fig. 11], reference constellation point shifter A334 (e.g., corresponding to method step M304 of Fig. 11], uplink transmission performer A335 (e.g., corresponding to method step M305 of Fig. 11], A functional unit may for instance correspond to a code block within a memory A310, A320, A340. At least one or more of functional units A331, A332, A333, A334, A335 may for instance be connected to and / or control and / or be part of the communication interface A350. The communication interface A350 may for instance comprise a power amplifier.

[0276] Fig. 17 shows an example block diagram of an apparatus 200 (e.g., performing and / or controlling a method according to the fourth example aspect, e.g., a network node]. The apparatus 200 may perform a methodaccording to the fourth example aspect. The apparatus 200 may comprise a user interface A460, a program memory A410, a main memory A420, and a data memory A440. Further, it comprises a processor A430. The apparatus 200 may further comprise functional units RF requirement parameter provider A431 (e.g., corresponding to method step M401 of Fig. 12], constellation request provider A432 (e.g., corresponding to method step M402 of Fig. 12], constellation distortion array indication obtainer A433 (e.g., corresponding to method step M403 of Fig. 12], A functional unit may for instance correspond to a code block within a memory A410, A420, A440. The functional units A431, A432, A433 may for instance be connected to and / or control the communication interface A450.

[0277] Fig. 18 shows an example block diagram of an apparatus 201 (e.g., performing and / or controlling a method according to the fifth example aspect, e.g., a network node]. The apparatus 201 may perform a method according to the fifth example aspect. The apparatus 201 may comprise a user interface A560, a program memory A510, amain memory A520, and a data memory A540. Further, it comprises a processor A530. The apparatus 201 may further comprise functional units calibration signal obtainer A531 (e.g., corresponding to method step M501 of Fig. 13], constellation distortion array determiner A532 (e.g., corresponding to method step M502 of Fig. 13], constellation distortion array indication provider A533 (e.g., corresponding to method step M503 of Fig. 13], A functional unit may for instance correspond to a code block within a memory A510, A520, A540. The functional units A531, A533 may for instance be connected to and / or control the communication interface A550.

[0278] Fig. 19 is a schematic illustration of examples of tangible and non-transitory computer-readable storage media according to the present invention that may for instance be used to implement program and / or main memory A110, A120, A140, A210, A220, A240, A310, A320, A340, A410, A420, A440, A510, A520, A540 of the apparatus 100, 101, 102 and / or 200, 201 of Fig. 9 to 13. Fig. 19 shows a flash memory 1900, which may for instance be soldered or bonded to a printed circuit board, a solid-state drive 1901 comprising a plurality of memory chips (e.g. Flash memory chips], a magnetic hard drive 1902, a Secure Digital (SD] card 1903, a Universal Serial Bus (USB] memory stick 1904, an optical storage medium 1905 (such as for instance a CD-ROM or DVD] and a magnetic storage medium 1906.

[0279] Any presented connection in the described embodiments is to be understood in a way that the involved components are operationally coupled. Thus, the connections can be direct or indirect with any number or combination of intervening elements, and there may be merely a functional relationship between the components.

[0280] Further, as used in this text, the term ‘circuitry’ refers to any of the following:

[0281] (a] hardware-only circuit implementations (such as implementations in only analog and / or digital circuitry]

[0282] (b] combinations of circuits and software (and / or firmware], such as: (i] to a combination of processors] or (ii] to sections of processor^] / software (including digital signal processor(s]],software, and memory(ies] that work together to cause an apparatus, such as a mobile phone, to perform various functions] and

[0283] (c] to circuits, such as a microprocessor(s] or a section of a microprocessor(s], that re-quire software or firmware for operation, even if the software or firmware is not physically present.

[0284] This definition of ‘circuitry’ applies to all uses of this term in this text, including in any claims. As a further example, as used in this text, the term ‘circuitry’ also covers an implementation of merely a processor (or multiple processors] or section of a processor and its (or their] accompanying software and / or firmware. The term ‘circuitry’ also covers, for example, a baseband integrated circuit or applications processor integrated circuit for a mobile phone.

[0285] Any of the processors mentioned in this text, in particular but not limited to processors 130, 230 of Figs. 7 and 8, could be a processor of any suitable type. Any processor may comprise but is not limited to one or more microprocessors, one or more processor(s] with accompanying digital signal processor(s], one or more processor(s] without accompanying digital signal processor(s], one or more special-purpose computer chips, one or more field-programmable gate arrays (FPGAS], one or more controllers, one or more applicationspecific integrated circuits (ASICS], or one or more computers], The relevant structure / hardware has been programmed in such a way to carry out the described function.

[0286] Moreover, any of the actions or steps described or illustrated herein may be implemented using executable instructions in a general-purpose or special-purpose processor and stored on a computer-readable storage medium (e.g., disk, memory, or the like] to be executed by such a processor. References to ‘computer-readable storage medium’ should be understood to encompass specialized circuits such as FPGAs, ASICs, signal processing devices, and other devices.

[0287] Moreover, any of the actions described or illustrated herein may be implemented using executable instructions in a general-purpose or special-purpose processor and stored on a computer-readable storage medium (e.g., disk, memory, or the like] to be executed by such a processor. References to ‘computer-readable storage medium’ should be understood to encompass specialized circuits such as FPGAs, ASICs, signal processing devices, and other devices.

[0288] The wording "A, or B, or C, or a combination thereof’ or "at least one of A, B and C” may be understood to be not exhaustive and to include at least the following: (i] A, or (ii] B, or (iii] C, or (iv] A and B, or (v] A and C, or (vi] B and C, or (vii] A and B and C.

[0289] It will be understood that the embodiments disclosed herein are only exemplary, and that any feature presented for a particular exemplary embodiment may be used with any aspect of the present disclosure on its own or in combination with any feature presented for the same or another particular exemplary embodiment and / or in combination with any other feature not mentioned. It will further be understood thatany feature presented for an example embodiment in a particular category may also be used in a corresponding manner in an example embodiment of any other category.

[0290] LIST OF ABBREVIATIONS CDV constellation distortion vector

[0291] DPD digital predistortion

[0292] DPOD digital postdistortion

[0293] SEM spectral emission mask

[0294] ACLR adjacent channel leakage ratio

Claims

C l a i m sA terminal device, comprising:means for providing a calibration signal to a network node using a reference constellation; and means for obtaining an indication of a constellation distortion array from the network node in response to providing the calibration signal, wherein a respective entry of the constellation distortion array indicates a shift in an In-phase / Quadrature, I / Q, plane of a respective reference constellation point of the reference constellation into a corresponding custom constellation point of a custom constellation.The terminal device according to claim 1, wherein the shift is specified in Cartesian coordinates or in Polar coordinates.The terminal device according to claim 1 or 2, wherein at least one ofthe reference constellation comprises uniformly distributed reference constellation points;the reference constellation corresponds to a square Quadrature Amplitude Modulation, QAM, constellation; orthe reference constellation corresponds to another custom constellation.The terminal device according to any of claims 1 to 3, wherein at least one ofthe custom constellation comprises unevenly distributed custom constellation points; or the custom constellation corresponds to a learned constellation.The terminal device according to any of claims 1 to 4, wherein the constellation distortion array indicates at least two mutually different shifts for at least two respective different reference constellation points of the reference constellation.The terminal device according to any of claims 1 to 5, wherein the constellation distortion array indicates respective shifts only for a subset of all the reference constellation points of the reference constellation.The terminal device according to any of claims 1 to 6, wherein the indication of the constellation distortion array corresponds to the constellation distortion array.The terminal device according to any of claims 1 to 7, further comprisingmeans for obtaining an indication of a calibration signal configuration from the network node; and means for configuring at least one property of the calibration signal based on the obtained calibration signal configuration.The terminal device according to claim 8, wherein at least one ofthe calibration signal configuration indicates a number of reference constellation points of the reference constellation to be used for the calibration signal;the calibration signal configuration indicates a type of a constellation to be used for the calibration signal; orthe calibration signal configuration indicates a transmission power to be used by the terminal device when providing the calibration signal.The terminal device according to any of claims 1 to 9, wherein the constellation distortion array is determined based on at least one RF requirement parameter.The terminal device according to claim 10, wherein the at least one RF requirement parameter comprises at least one ofan out-of-band interference requirement parameter;an in-band interference requirement parameter; ora spectral mask transmission requirement parameter.The terminal device according to any of claims 1 to 11, further comprising means for providing a constellation distortion capability indication to the network node, wherein the constellation distortion capability indicates a capability to utilize a constellation distortion array for transforming a constellation.The terminal device according to any of claims 1 to 12, further comprising means for performing at least one uplink transmission to the network node, wherein the at least one uplink transmission is modulated using the custom constellation.The terminal device according to any of claim 13, wherein the modulating comprises applying a normalization factor to the custom constellation, wherein the normalization factor is based on the constellation distortion array.A network node, comprisingmeans for obtaining a calibration signal from a terminal device based on a reference constellation; means for determining a constellation distortion array at least partially based on the obtained calibration signal, wherein a respective entry of the constellation distortion array indicates a shift in an In-phase / Quadrature, I / Q, plane of a respective reference constellation point of the reference constellation into a corresponding custom constellation point of a custom constellation; and means for providing an indication of the determined constellation distortion array to the terminal device.The network node according to claim 15, wherein the shift is specified in Cartesian coordinates or in Polar coordinates.The network node according to claim 15 or 16, wherein at least one ofthe reference constellation comprises uniformly distributed reference constellation points;the reference constellation corresponds to a square Quadrature Amplitude Modulation, QAM, constellation; orthe reference constellation corresponds to another custom constellation.The network node according to any of claims 15 to 17, wherein at least one ofthe custom constellation comprises unevenly distributed custom constellation points; or the custom constellation corresponds to a learned constellation.The network node according to any of claims 15 to 18, wherein the constellation distortion array indicates at least two mutually different shifts for at least two respective different reference constellation points of the reference constellation.The network node according to any of claims 15 to 19, wherein the constellation distortion array indicates respective shifts only for a subset of all the reference constellation points of the reference constellation.The network node according to any of claims any of claims 15 to 20, wherein the indication of the constellation distortion array corresponds to the constellation distortion array.The network node according to any of claims 15 to 21, further comprisingmeans for providing an indication of a calibration signal configuration to the terminal device, wherein the calibration signal configuration indicates at least one property of the calibration signal.The network node according to claim 22, wherein at least one ofthe calibration signal configuration indicates a number of reference constellation points of the reference constellation to be used for the calibration signal;the calibration signal configuration indicates a type of a constellation to be used for the calibration signal; orthe calibration signal configuration indicates a transmission power to be used by the terminal device when providing the calibration signal.The network node according to any of claims 15 to 23, further comprising of means for obtaining a constellation distortion capability indication from the terminal device, wherein the constellation distortion capability indicates a capability to utilize a constellation distortion array for transforming a constellation.The network node according to any of claims any of claims 15 to 24, wherein the determining the constellation distortion array is further based on at least one radio frequency, RF, requirement.The network node according to claim 25, wherein the at least one RF requirement parameter comprises at least one ofan out-of-band interference requirement parameter;an in-band interference requirement parameter; ora spectral mask transmission requirement parameter.The network node according to any of claims any of claims 15 to 26, wherein the determining the constellation distortion array comprises at least one ofgenerating a constellation distortion array; orselecting a constellation distortion array.The network node according to claim 27, wherein selecting the constellation distortion array comprises selecting the constellation distortion array from a table using the at least one RF requirement parameter.The network node according to claim 28, wherein at least one ofthe table is specific to the terminal device or to a class of the terminal device; orthe table is specific to a transmission power.The network node according to any of claims any of claims 15 to 29, further comprising means for adjusting a receiver of the network node in accordance with the determined constellation distortion array.The network node according to any of claims 15 to 30, further comprisingmeans for receiving at least one uplink transmission from the terminal device and demodulating the at least one uplink transmission using the custom constellation.