Repeated and offset QPSK DFT-s-OFDM modulation

Abstract

The invention relates to a transmitter device (100) and a receiver device (300) for repeated and offset QPSK DFT-s-OFDM modulation. The transmitter device (100) is configured to split a stream of bits (130) into an in-phase stream of bits (132) and a quadrature stream of bits (134) which are encoded with a 1 / R rate repetition code into an encoded in-phase stream of bits (142) and an encoded quadrature stream of bits (144) such that the encoded quadrature stream of bits (144) is offset in relation to the encoded in-phase stream of bits (142) with F number of bits, or vice versa. The encoded in-phase stream of bits (142) and the encoded quadrature stream of bits (144) are jointly modulated into a stream of QPSK symbols (150) which are modulated into DFT-s-OFDM, symbols (160) for transmission to the receiver device (300). The invention relates to corresponding methods and a computer program.

Description

[0001]REPEATED AND OFFSET QPSK DFT-s-OFDM MODULATION TECHNICAL FIELDExamples of the invention relate to a transmitter device and a receiver device for repeated and offset QPSK DFT-s-OFDMmodulation for communication systems. Furthermore, examples of the invention also relate to corresponding methods and acomputer program. BACKGROUND Satellite communications is envisioned to be an important feature of next 3GPP releases for 5.5G and 6G, with true new market opportunities for the wireless industry. Satellite communications could provide wide-area coverage with ubiquitousconnectivity which may be needed enablers for some applications such as for example remote internet of things (IoT) andvehicle-to-everything (V2X). In previous studies, 3GPP showed that normal user equipment (UE) handsets (so-calledhandheld) could be used to connect both to terrestrial network (TN) and non-terrestrial network (NTN), and not only dedicated satellite UE with a higher power class (so-called VSAT) as with traditional satellite operators. Notably, 3GPP showed thatsatellite connectivity could be achieved with current 5G new radio (NR) waveform based on cyclic prefix orthogonal frequencydivision multiplexing (CP-OFDM) for downlink (DL) and discrete Fourier transform spread orthogonal frequency divisionmultiplexing (DFT-s-OFDM) for uplink (UL) without a dedicated satellite waveform, such that NR standard could provideconnectivity to both TN and NTN. This conclusion from 3GPP is implicitly tight to the low-data rate assumption and low-frequency band considered for NR NTN. Recent Rel.-18 work item on NTN evolution was targeting the evaluation of NTN coverage with already potential issues forproviding voice over IP (VoIP) and low-data rate services for commercial handset terminals. This is of very low-data ratecompared to for example IMT-2020 specification of maximum data rate for 5G being 20Gbps (downlink) and 10Gbps (uplink). In satellite communications, due to the very small link budget, both the UE and the satellite are very frequently transmitting close to saturation level of the power amplifier. Therefore, for traditional satellite operators low peak-to-average power ratio (PAPR) signals are of specific importance. Also because of the very small link budget, most traditional satellite operators use low-order modulations. Another difference with TN channel is that satellite channels are typically made of only few resolvable paths often in line-of-sign (LOS) conditions, and thus NTN channels has very little frequency selectivity and so most satellitechannel modelling relies on frequency flat fading. Finally, while satellite communication is characterized by high Doppler shiftand resulting large frequency offset, 3GPP showed that for UE with global navigation satellite system (GNSS) capabilities,pre-compensation in both in UL and DL can reduce greatly the frequency offset, and from this legacy NR signals can be usedfor frequency synchronization. The residual frequency offset for data transmission can thus be assumed very small such thatchannel is almost constant over several OFDM symbols. SUMMARYAn objective of examples of the invention is to provide a solution which mitigates or solves the drawbacks and problems ofconventional solutions. Another objective of examples of the invention is to provide a modulation scheme having lower PAPR compared to conventional solutions.Yet another objective of examples of the invention is to provide an improved error rate performance at low signal-to-noise ratio(SNR) compared to conventional solutions.The above and further objectives are solved by the subject matter of the independent claims. Further examples of the inventioncan be found in the dependent claims.According to a first aspect of the invention, the above mentioned and other objectives are achieved with a transmitter deviceconfigured to:split a stream of bits into an in-phase stream of bits and a quadrature stream of bits;encode the in-phase stream of bits and the quadrature stream of bits with a 1 / ^ rate repetition code into an encoded in-phase stream of bits and an encoded quadrature stream of bits, wherein the encoded quadrature stream of bits is offset in relationto the encoded in-phase stream of bits with ^ number of bits, or vice versa, where ^ and ^ are positive integers;jointly modulate the encoded in-phase stream of bits and the encoded quadrature stream of bits into a stream ofQuadrature Phase Shift Keying, QPSK, symbols;modulate the stream of QPSK symbols into one or more Discrete Fourier Transform spread Orthogonal Frequency Division Multiplexing, DFT-s-OFDM, symbols; and transmit the one or more DFT-s-OFDM symbols.The one or more DFT-s-OFDM symbols may be transmitted to a receiver device in a communication system.An advantage of the transmitter device according to the first aspect is that it enables a stream of QPSK symbols whoseconsecutive phase is never larger than ^ / 2 which limits envelope fluctuation and thus PAPR compared to conventionalsolutions.In an implementation form of a transmitter device according to the first aspect, ^ is dependent on ^.An advantage with this implementation form is that the offset is such that the repeated pattern of the I-branch and Q-branchare never aligned, and thus guaranteeing limited phase fluctuation and reduced PAPR.In an implementation form of a transmitter device according to the first aspect, 1 ≤ ^ ≤ ^ − 1 modulo ^.An advantage with this implementation form is that it guarantees that the offset is such that the repeated pattern of the I-branch and Q-branch are never aligned, and thus guaranteeing limited phase fluctuation and reduced PAPR.In an implementation form of a transmitter device according to the first aspect, ^ = 2.An advantage with this implementation form is that it maximizes the rate of the repetition and thus the spectral efficiency ofthe transmission. In an implementation form of a transmitter device according to the first aspect, the transmitter device is configured to: flip at least one bit in each segment of ^ number of encoded bits of the encoded in-phase stream of bits and flip at leastone bit in each segment of ^ number of encoded bits of the encoded quadrature stream of bits; oralternate a sign of at least one QPSK symbol in each segment of ^ number of QPSK symbols of the stream of QPSKsymbols.An advantage with this implementation form is that it mitigates an inherent phase difference among consecutive time-multiplexed DFT-s-OFDM pulses / symbols, thereby mitigating the phase fluctuation of the transmitted DFT-s-OFDM signal and thus reducing PAPR. In an implementation form of a transmitter device according to the first aspect, the modulating of the stream of QPSK symbolsinto the one or more DFT-s-OFDM symbols comprises:DFT modulate the stream of QPSK symbols into a stream of Fourier coefficients; Frequency Domain Spectral Shaping, FDSS, the stream of Fourier coefficients into a stream of shaped Fouriercoefficients based on a bell-shaped FDSS window; andInverse Fast Fourier Transform, IFFT, modulating the stream of shaped Fourier coefficients into the one or more DFT-s-OFDM symbols.An advantage with this implementation form is that it will improve the effective SNR of the transmission and thus theinformation rate of mentioned transmission. In an implementation form of a transmitter device according to the first aspect, the bell-shaped FDSS window is based on: where ^^^ is the number of subcarriers used for transmitting the one or more DFT-s-OFDM symbols, and ^ is the subcarrierindex.An advantage with this implementation form is that it will optimize the effective SNR of the transmission and thus theinformation rate of mentioned transmission.In an implementation form of a transmitter device according to the first aspect, the transmitter device is configured to:cyclic shift the stream of Fourier coefficients by ^^^ / 2. An advantage with this implementation form is that it maps the zero-frequency component outputted by the DFT precoding tothe direct current subcarrier of the OFDM transmission, which minimizes transmitted signal envelope fluctuation and thusPAPR.In an implementation form of a transmitter device according to the first aspect, the transmitter device is configured to:transmit the one or more DFT-s-OFDM symbols on an even number of subcarriers.An advantage with this implementation form is that it enables to use a ^ = 2 repetition code for all input bits which maximizesthe spectral efficiency of the transmission.In an implementation form of a transmitter device according to the first aspect, the stream of bits is a stream of information bits encoded with a Forward Error Correction, FEC, code.An advantage with this implementation form is that it enables error correction at the receiver device reaching higher spectralefficiency of the transmission.In an implementation form of a transmitter device according to the first aspect, the transmitter device is configured to: transmit a control message to a receiver device, the control message indicating ^ and / or ^. An advantage with this implementation form is that it enables the transmitter device to choose the most suitable parameterswhile letting the receiver device to know the repetition coding and / or the offset of the transmitted stream of QPSK symbols,thereby allowing the combining at the receiver device providing higher effective SNR and better performance. According to a second aspect of the invention, the above mentioned and other objectives are achieved with a receiver device configured to: receive one or more DFT-s-OFDM symbols; demodulate the one or more DFT-s-OFDM symbols into a stream of demodulated constellation symbols; successive combining ^ number of demodulated constellation symbols into a stream of combined symbols, where ^ isa positive integer; split a first half of the stream of combined symbols into a stream of in-phase symbols and a second half of the streamof combined symbols into a stream of quadrature symbols;jointly demodulate the stream of in-phase symbols and the stream of quadrature symbols into a stream of soft QPSKsymbols; anddecode the stream of soft QPSK symbols into a stream of decoded bits.The one or more DFT-s-OFDM symbols may be received from a transmitter device in a communication system.An advantage of the receiver device according to the second aspect is that the receiver device appropriately combines the streamof QPSK symbols before demodulation and decoding according to the modulation scheme at the transmitter device, therebyproviding higher effective SNR and better performance.In an implementation form of a receiver device according to the second aspect, the successive combining of the ^ number ofdemodulated constellation symbols into the stream of combined symbols comprises: average the ^ number of demodulated constellation symbols with an alternating sign change.An advantage with this implementation form is that the receiver device inverts alternating sign changes that was made at thetransmitter device leading to constructive combining and improved SNR and error rate performance.In an implementation form of a receiver device according to the second aspect, the splitting of the first half of the stream of combined symbols into the stream of in-phase symbols and the second half of the stream of combined symbols into the streamof quadrature symbols is based on an offset of ^ number of combined symbols, where ^ is positive integer.An advantage with this implementation form is that the receiver device takes into account the offset made at the transmitterdevice when combining leading to improved SNR and error rate performance. An advantage with this implementation form is that it takes into account the effect of the combiner on the effective SNR in demodulation, which improves the error rate performance. In an implementation form of a receiver device according to the second aspect, the receiver device is configured to: receive the one or more DFT-s-OFDM symbols on an even number of subcarriers.An advantage with this implementation form is that it enables a pairwise combining of symbols with rate ½ repetition codingat the transmitter device thereby providing the best spectral efficiency of the transmission.In an implementation form of a receiver device according to the second aspect, the receiver device is configured to: receive a control message from a transmitter device, the control message indicating ^ and / or ^.An advantage with this implementation form is that it allows the receiver device to know the corresponding transmissionparameters used by the transmitter device and may therefore perform the demodulation accordingly.In an implementation form of a receiver device according to the second aspect, ^ and / or ^ are predetermined.An advantage with this implementation form is that it limits signaling overhead in the communication system.According to a third aspect of the invention, the above mentioned and other objectives are achieved with a method for a transmitter device, the method comprises: splitting a stream of bits into an in-phase stream of bits and a quadrature stream of bits; encoding the in-phase stream of bits and the quadrature stream of bits with a 1 / ^ rate repetition code into an encodedin-phase stream of bits and an encoded quadrature stream of bits, wherein the encoded quadrature stream of bits is offset inrelation to the encoded in-phase stream of bits with ^ number of bits, or vice versa, where ^ and ^ are positive integers;jointly modulating the encoded in-phase stream of bits and the encoded quadrature stream of bits into a stream of QPSKsymbols; modulating the stream of QPSK symbols into one or more DFT-s-OFDM symbols; and transmitting the one or more DFT-s-OFDM symbols.The method according to the third aspect can be extended into implementation forms corresponding to the implementationforms of the transmitter device according to the first aspect. Hence, an implementation form of the method comprises thefeature(s) of the corresponding implementation form of the transmitter device. The advantages of the methods according to the third aspect are the same as those for the corresponding implementation formsof the transmitter device according to the first aspect.According to a fourth aspect of the invention, the above mentioned and other objectives are achieved with a method for a receiver device, the method comprises: receiving one or more DFT-s-OFDM symbols; demodulating the one or more DFT-s-OFDM symbols into a stream of demodulated constellation symbols;successive combining ^ number of demodulated constellation symbols into a stream of combined symbols, where ^ isa positive integer; splitting a first half of the stream of combined symbols into a stream of in-phase symbols and a second half of the stream of combined symbols into a stream of quadrature symbols; jointly demodulating the stream of in-phase symbols and the stream of quadrature symbols into a stream of soft QPSK symbols; and decoding the stream of soft QPSK symbols into a stream of decoded bits. The method according to the fourth aspect can be extended into implementation forms corresponding to the implementationforms of the receiver device according to the second aspect. Hence, an implementation form of the method comprises thefeature(s) of the corresponding implementation form of the receiver device.The advantages of the methods according to the fourth aspect are the same as those for the corresponding implementation formsof the receiver device according to the second aspect.Examples of the invention also relate to a computer program, characterized in program code, which when run by at least oneprocessor causes the at least one processor to execute any method according to examples of the invention. Further, examplesof the invention also relate to a computer program product comprising a computer readable medium and the mentioned computer program, wherein the computer program is included in the computer readable medium, and may comprises one ormore from the group of: read-only memory (ROM), programmable ROM (PROM), erasable PROM (EPROM), flash memory,electrically erasable PROM (EEPROM), hard disk drive, etc.Further applications and advantages of examples of the invention will be apparent from the following detailed description.BRIEF DESCRIPTION OF THE DRAWINGSThe appended drawings are intended to clarify and explain different examples of the invention, in which:^ Fig. 1 shows a transmitter device according to examples of the invention;^Fig. 2 shows a flow chart of a method for a transmitter device according to examples of the invention;^Fig. 3 shows a receiver device according to examples of the invention;^ Fig. 4 shows a flow chart of a method for a receiver device according to examples of the invention;^ Fig. 5 shows a communication system according to examples of the invention; and^ Fig. 6 illustrates and 7a and 7b illustrates different encoding aspects according to examples of the invention;^Fig. 8 shows a transmitter device and a receiver device according to examples of the invention; and^Figs. 9 – 11 show performance results of examples of the invention. The signal is received via a multi-tap channel for indices 0 ≤ ^ ≤ (^^^^ − 1) as (3)The time-domain noise {^[^]} is assumed to be zero-mean with unit average energy, i.e. ^[|^[^]|^] = 1.OFDM demodulation with one-tap equalization is based on the assumption that the channel taps are constant over each OFDMsymbol, i.e., ℎ^[^] = ℎ^, and the CP is longer than the maximum delay spread. Under these assumptions, the receivedfrequency-domain (FD) symbol on subcarrier ^ after FFT demodulation can be written as^[^] = ^[^] ^[^] + ^[^] (4)with per-subcarrier channel coefficient ^[^] = is the unnormalized discreteFourier transform (DFT) of the channel impulse response at subcarrier ^. The frequency-domain noise ^[^] has also unitvariance as ^[| ^[^]|^] = 1.The subcarrier-level symbols are then equalized as ^[^] = ^[^]^[^] = ^[^]^[^] + ^[^]^[^]. (5)After equalization, the inverse DFT precoding is applied on the equalized symbols ^^[^], from which the received symbols are given as ^[^] = ^ ^[^] + ICI[^] + ^[^] (6)where the useful channel component is independent of ^ and given by The intercarrier interference from the equalizer is given as with interference channels and post-processed noise term Treating the interference as noise and assuming the noise being Gaussian, the effective signal to noise ratio (SNR) of thetransmission is in general where ^ is given in Eq. (7) and assumed real as result of equalization. The average total power from the signal components is and the noise power is Given the channel state ^[^], typical equalizers include match filter (MF), zero-forcing (ZF), and minimum mean square error (MMSE), defined respectively as For ZF there is no inter channel interference (ICI), and one gets, ^ = 1, P^^^ = 1, P^^^ − ^^ = 0, and the effective SNR is For MMSE, the noise-plus-interference power simplifies to (P^^^ − ^^ + ^^^) = (1 − ^)^ and thus^^^,^ (16) SNR^^^^=(1 − ^) When SNR → 0, then MMSE → MF, while when SNR → ∞, then MMSE → ZF. Here we are interested low-order modulationoperating in low SNR for which MMSE and MF perform closely, but in general, MMSE provides the best performance with DFT-s-OFDM. Given a bit ^^, we will write the bit to amplitude mapping as So, ^^ ^ ^= 0 is mapped to amplitude ^^ = √^, and ^^ = 1 is mapped to the opposite direction ^^ = − √^. Given two consecutive bits ^^and ^^^^, a Gray-mapped QPSK symbol is given by ^^^^^(^^ , ^^^^) = ^^ + ^^^^^ , (18)i.e., the modulation is based on a direct serial-to-parallel conversion to in-phase (I) branch and quadrature (Q) branch asillustrated in Table 1 below when modulating bits to the ^^^ subcarriers. In this table, the stream of input bits is converted fromserial to parallel, by populating the I-branch and Q-branch of the first subcarrier index, next the I-branch and Q-branch of thesecond subcarrier index, and so one. The QPSK modulation is the mapping of each of these bits on the I-branch and Q-branchto the amplitude values Table 1: Illustration of QPSK modulation QPSK Input Bits Constellation symbolsI-branch Q-branch I-branch Q-branch^^^^^^^^(^^ , ^^ , … . , ^^^^^^^)^^^^^^^^…. …. …. ….^^(^^^^^)^^^^^^^^^(^^^^^)^^^^^^^Given the received signal ^[^] in Eq. (6), the soft-symbol from the I-branch and Q-branch can be detected as^^ Re{^[^]} ^ even^ ≈ ^(19) Im{^[^]} ^ odd. Given a bit ^^, a BSPK constellation symbol is obtained as BPSK can be seen as QPSK symbol where bits are encoded with 1 / 2-rate repetition coding before serial-to-parallel conversion.This is illustrated in Table 2 below where we see the stream of input bits is mapped to the I-branch and repeated on the Q-branch, or vice-versa, and then QPSK modulation is applied to generate the BPSK symbols.Table 2: Illustration of BPSK modulation BPSK Input Bits Constellation symbolsI-branch Q-branch I-branch Q-branch^^^^^^^^(^^ , ^^ , … . , ^^^^^^)^^^^^^^^…. …. …. ….^^^^^^^^^^^^^^^^^^^^^^^^Given the received signal ^[^], the soft- the symbols to the I-branch as (21) Given bit ^^the constellation symbol is given where ^^^= 1 ⊕ ^^(mod 2) is a bit flipping operation. Here ^ / 2-BPSK can be seen as QPSK symbol where bits are encodedwith 1 / 2-rate coding with bit-flip repetition as illustrated in Table 3 below where ^^^ is assumed even. In this table, we see thatas for BPSK the stream of input bits is mapped to the I-branch and repeated on the Q-branch. However, here every second biton the I-branch is flipped. Then QPSK modulation is applied to generate the π / 2-BPSK symbols. The bit-flipping operation results into a sign change on the I-branch of the constellation symbols. Table 3: Illustration of ^ / ^-BPSK modulation ^ / 2-BPSK Input Bits Constellation symbolsI-branch Q-branch I-branch Q-branch ^^^^^^^^(^^ , ^^ , … . , ^^^^^^)^^^ ^^ −^^ ^^^^^^^^^^…. …. …. ….^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ ^^^^^^ −^^^^^^ ^^^^^^ An objective of examples of the invention is to disclose a new modulation scheme based on QPSK DFT-s-OFDM modulationwith low PAPR and improved error performance and that may operate in low SNR region and thus particularly relevant butnot limited to satellite communications. This is especially the case for lowest modulation and coding scheme (MCS) indices ofNR uplink. The novel modulation scheme is directly compatible with legacy DFT-s-OFDM transmitter devices and receiverdevices and legacy channel codes without introducing any new complex signal processing.Fig. 1 shows a transmitter device 100 according to an example of the invention. In the example shown in Fig. 1, the transmitterdevice 100 comprises a processor 102, a transceiver 104 and a memory 106. The processor 102 is coupled to the transceiver104 and the memory 106 by communication means 108 known in the art. The transmitter device 100 may be configured forwireless and / or wired communications in a communication system. The wireless communication capability may be provided with an antenna or antenna array 110 coupled to the transceiver 104, while the wired communication capability may be provided with a wired communication interface 112 e.g., coupled to the transceiver 104. The processor 102 may be referred to as one or more general-purpose central processing units (CPUs), one or more digitalsignal processors (DSPs), one or more application-specific integrated circuits (ASICs), one or more field programmable gatearrays (FPGAs), one or more programmable logic devices, one or more discrete gates, one or more transistor logic devices, one or more discrete hardware components, or one or more chipsets. The memory 106 may be a read-only memory, a randomaccess memory (RAM), or a non-volatile RAM (NVRAM). The transceiver 104 may be a transceiver circuit, a power controller,or an interface providing capability to communicate with other communication modules or communication devices, such asnetwork nodes and network servers. The transceiver 104, memory 106 and / or processor 102 may be implemented in separatechipsets or may be implemented in a common chipset. That the transmitter device 100 is configured to perform certain actionscan in this disclosure be understood to mean that the transmitter device 100 comprises suitable means, such as e.g., the processor102 and the transceiver 104, configured to perform the actions.According to examples of the invention the transmitter device 100 is configured to split a stream of bits 130 into an in-phasestream of bits 132 and a quadrature stream of bits 134; encode the in-phase stream of bits 132 and the quadrature stream of bits134 with a 1 / ^ rate repetition code into an encoded in-phase stream of bits 142 and an encoded quadrature stream of bits 144,wherein the encoded quadrature stream of bits 144 is offset in relation to the encoded in-phase stream of bits 142 with ^ numberof bits, or vice versa, where ^ and ^ are positive integers; jointly modulate the encoded in-phase stream of bits 142 and theencoded quadrature stream of bits 144 into a stream of QPSK symbols 150; modulate the stream of QPSK symbols 150 intoone or more DFT-s-OFDM symbols 160; and transmit the one or more DFT-s-OFDM symbols 160.Furthermore, in an example of the invention, the transmitter device 100 comprises processor configured to: split a stream ofbits 130 into an in-phase stream of bits 132 and a quadrature stream of bits 134; encode the in-phase stream of bits 132 and thequadrature stream of bits 134 with a 1 / ^ rate repetition code into an encoded in-phase stream of bits 142 and an encodedquadrature stream of bits 144, wherein the encoded quadrature stream of bits 144 is offset in relation to the encoded in-phasestream of bits 142 with ^ number of bits, or vice versa, where ^ and ^ are positive integers; jointly modulate the encoded in-phase stream of bits 142 and the encoded quadrature stream of bits 144 into a stream of QPSK symbols 150; modulate thestream of QPSK symbols 150 into one or more DFT-s-OFDM symbols 160. The transmitter device 100 comprises a transceiverconfigured to: transmit the one or more DFT-s-OFDM symbols 160.Moreover, in yet another example of the invention, the first communication 100 for a communication system 500 comprises aprocessor and a memory having computer readable instructions stored thereon which, when executed by the processor, causethe processor to: split a stream of bits 130 into an in-phase stream of bits 132 and a quadrature stream of bits 134; encode thein-phase stream of bits 132 and the quadrature stream of bits 134 with a 1 / ^ rate repetition code into an encoded in-phasestream of bits 142 and an encoded quadrature stream of bits 144, wherein the encoded quadrature stream of bits 144 is offsetin relation to the encoded in-phase stream of bits 142 with ^ number of bits, or vice versa, where ^ and ^ are positive integers;jointly modulate the encoded in-phase stream of bits 142 and the encoded quadrature stream of bits 144 into a stream of QPSKsymbols 150; modulate the stream of QPSK symbols 150 into one or more DFT-s-OFDM symbols 160; and transmit the oneor more DFT-s-OFDM symbols 160.Fig. 2 shows a flow chart of a corresponding method 200 which may be executed in a transmitter device 100, such as the oneshown in Fig. 1. The method 200 comprises: splitting 202 a stream of bits 130 into an in-phase stream of bits 132 and aquadrature stream of bits 134; encoding 204 the in-phase stream of bits 132 and the quadrature stream of bits 134 with a 1 / ^ rate repetition code into an encoded in-phase stream of bits 142 and an encoded quadrature stream of bits 144, wherein the encoded quadrature stream of bits 144 is offset in relation to the encoded in-phase stream of bits 142 with ^ number of bits, orvice versa, where ^ and ^ are positive integers; jointly modulating 206 the encoded in-phase stream of bits 142 and the encodedquadrature stream of bits 144 into a stream of QPSK symbols 150; modulating 208 the stream of QPSK symbols 150 into oneor more DFT-s-OFDM symbols 160; and transmitting 210 the one or more DFT-s-OFDM symbols 160.To split the stream of bits 130 into the in-phase stream of bits 132 and the quadrature stream of bits 134 may be understood assuch that each bit in the stream of bits is allocated either to the I-branch or Q-branch. For example, for a set of bits ℬ, it ispartitioned as ℬ = ℬ^ ∪ ℬ^ in two non-intersecting subsets ℬ^ and ℬ^, i.e. ℬ^ ∩ ℬ^ = ∅ , and each subset is mapped to theI-branch and Q-branch, respectively. ^ and ^ are positive integers and if there is no repetition and no offset ^ = 1 and ^ = 0, which means that the transmissionis a QPSK modulation as in conventional solutions. The limited PAPR effect can be obtained using ^ = 2 but larger ^ than^ = 2 may be needed to accommodate different system configuration such as odd number of subcarrier allocation, whichwould provide similar PAPR effect at the cost of lower spectral efficiency. In general, due cyclicity the offset value ^ can belimited to be less than the number of subcarriers employed in the transmission.To jointly modulate the encoded in-phase stream of bits 142 and the encoded quadrature stream of bits 144 into a stream ofQPSK symbols 150 may be understood as that each QPSK symbol is constructed as in Eq. (18) where the first input is takenfrom encoded in-phase stream of bits 142 and the second input is the encoded quadrature stream of bits 144.Moreover, according to examples of the invention ^ is dependent on ^ as multiple offset values ^ would provide an equivalenteffect depending on ^. Notably it is enough to restrict the offset value ^ to be in the range 1 ≤ ^ ≤ ^ − 1 modulo ^. Specially,^ = 2 provides the best trade-off performance between PAPR limitation and maximal spectral efficiency, for which the offsetcan be set to be ^ = 1.Fig. 3 shows a receiver device 300 according to an example of the invention. In the example shown in Fig. 3, the receiverdevice 300 comprises a processor 302, a transceiver 304 and a memory 306. The processor 302 is coupled to the transceiver304 and the memory 306 by communication means 308 known in the art. The receiver device 300 may be configured for wireless and / or wired communications in a communication system. The wireless communication capability may be providedwith an antenna or antenna array 310 coupled to the transceiver 304, while the wired communication capability may be providedwith a wired communication interface 312 e.g., coupled to the transceiver 304. The processor 302 may be referred to as one or more general-purpose CPUs, one or more DSPs, one or more ASICs, one or more FPGAs, one or more programmable logic devices, one or more discrete gates, one or more transistor logic devices, oneor more discrete hardware components, one or more chipsets. The memory 306 may be a read-only memory, a RAM, or aNVRAM. The transceiver 304 may be a transceiver circuit, a power controller, or an interface providing capability tocommunicate with other communication modules or communication devices. The transceiver 304, the memory 306 and / or theprocessor 302 may be implemented in separate chipsets or may be implemented in a common chipset. That the receiver device300 is configured to perform certain actions can in this disclosure be understood to mean that the receiver device 300 comprisessuitable means, such as e.g., the processor 302 and the transceiver 304, configured to perform the actions.According to examples of the invention the receiver device 300 is configured to receive one or more DFT-s-OFDM symbols160; demodulate the one or more DFT-s-OFDM symbols 160 into a stream of demodulated constellation symbols 330;successive combining ^ number of demodulated constellation symbols 330 into a stream of combined symbols 332, where ^is a positive integer; split a first half of the stream of combined symbols 332 into a stream of in-phase symbols 342 and a second half of the stream of combined symbols 332 into a stream of quadrature symbols 344; jointly demodulate the stream of in-phase symbols 342 and the stream of quadrature symbols 344 into a stream of soft QPSK symbols 350; and decode the streamof soft QPSK symbols 350 into a stream of decoded bits 352.From the perspective of the receiver device 300 ^ is the number of QPSK symbols in which a bit is carried / transmitted. Ingeneral, the receiver device 300 must know the modulation format so as to be able to decode the transmission from the transmitter device 100. The modulation format herein disclosed is based on parameter ^.Furthermore, in an example of the invention, the receiver device 300 comprises a transceiver configured to: receive one ormore DFT-s-OFDM symbols 160. The receiver device 300 comprises a processor configured to: demodulate the one or moreDFT-s-OFDM symbols 160 into a stream of demodulated constellation symbols 330; successive combining ^ number ofdemodulated constellation symbols 330 into a stream of combined symbols 332, where ^ is a positive integer; split a first halfof the stream of combined symbols 332 into a stream of in-phase symbols 342 and a second half of the stream of combined symbols 332 into a stream of quadrature symbols 344; jointly demodulate the stream of in-phase symbols 342 and the stream of quadrature symbols 344 into a stream of soft QPSK symbols 350; and decode the stream of soft QPSK symbols 350 into a stream of decoded bits 352Moreover, in yet another example of the invention, the receiver device 300 for a communication system 500 comprises aprocessor and a memory having computer readable instructions stored thereon which, when executed by the processor, causethe processor to: receive one or more DFT-s-OFDM symbols 160; demodulate the one or more DFT-s-OFDM symbols 160into a stream of demodulated constellation symbols 330; successive combining ^ number of demodulated constellation symbols330 into a stream of combined symbols 332, where ^ is a positive integer; split a first half of the stream of combined symbols332 into a stream of in-phase symbols 342 and a second half of the stream of combined symbols 332 into a stream of quadraturesymbols 344; jointly demodulate the stream of in-phase symbols 342 and the stream of quadrature symbols 344 into a streamof soft QPSK symbols 350; and decode the stream of soft QPSK symbols 350 into a stream of decoded bits 352.Fig. 4 shows a flow chart of a corresponding method 400 which may be executed in a receiver device 300, such as the oneshown in Fig. 3. The method 400 comprises: receiving 402 one or more DFT-s-OFDM symbols 160; demodulating 404 theone or more DFT-s-OFDM symbols 160 into a stream of demodulated constellation symbols 330; successive combining 406^ number of demodulated constellation symbols 330 into a stream of combined symbols 332, where ^ is a positive integer;splitting 408 a first half of the stream of combined symbols 332 into a stream of in-phase symbols 342 and a second half of the stream of combined symbols 332 into a stream of quadrature symbols 344; jointly demodulating 410 the stream of in-phasesymbols 342 and the stream of quadrature symbols 344 into a stream of soft QPSK symbols 350; and decoding 412 the streamof soft QPSK symbols 350 into a stream of decoded bits 352.Further details related to examples of the invention will now be described in a 3GPP 5G context. Thus, 3GPP 5G terminology, definitions, expressions and system architecture may be used.Fig. 5 shows a communication system 500 according to an example of the invention. The communication system 500 in thedisclosed example comprises a transmitter device 100 and a receiver device 300 configured to communicate and operate in thecommunication system 500. For simplicity, the shown communication system 500 only comprises one transmitter device 100and one receiver device 300. However, the communication system 500 may comprise any number of transmitter devices 100and any number of receiver devices 300 without deviating from the scope of the invention. The communication system 500may be a 3GPP communication system.In the disclosed example, the transmitter devices 100 act as a client device such as a UE. The receiver device 300 on the otherhand act as a network access node such as a base station or a satellite of a communication system. The network access nodemay be in communication with a network (NW) such as a core network of a 3GPP system. However, the reverse example isalso possible, i.e., that the receiver device 300 is acting as a client device while the transmitter devices 100 is acting as a networkaccess node or a satellite. The network access node may also be denoted as a radio network access node, an access network access node, an access point (AP), or a base station (BS), e.g., a radio base station (RBS), which in some networks may be referred to as transmitter, “gNB”, “gNodeB”, “eNB”, “eNodeB”, “NodeB” or “B node”, depending on the standard, technology and terminology used. The radio network access node may be of different classes or types such as e.g., macro eNodeB, home eNodeB or pico base station, basedon transmission power and thereby the cell size. The radio network access node may further be a station, which is any devicethat contains an IEEE 802.11-conformant media access control (MAC) and physical layer (PHY) interface to the wirelessmedium (WM). The radio network access node may be configured for communication in 3GPP related long term evolution(LTE), LTE-advanced, fifth generation (5G) wireless systems, such as new radio (NR) and their evolutions, as well as in IEEErelated Wi-Fi, worldwide interoperability for microwave access (WiMAX) and their evolutions.The client device may be denoted as a user device, a user equipment (UE), a mobile station, an internet of things (IoT) device, a sensor device, a wireless terminal and / or a mobile terminal, and is enabled to communicate wirelessly in a wireless communication system, sometimes also referred to as a cellular radio system. The UEs may further be referred to as mobile telephones, cellular telephones, computer tablets or laptops with wireless capability. The UEs in this context may be, for example, portable, pocket-storable, hand-held, computer-comprised, or vehicle-mounted mobile devices, enabled to communicate voice and / or data, via a radio access network (RAN), with another communication entity, such as another receiveror a server. The UE may further be a station, which is any device that contains an IEEE 802.11-conformant MAC and PHYinterface to the WM. The UE may be configured for communication in 3GPP related LTE, LTE-advanced, 5G wireless systems,such as NR, and their evolutions, as well as in IEEE related Wi-Fi, WiMAX and their evolutions.In examples of the invention, the transmitter device 100 may be configured to transmit a control message 510 to the receiverdevice 300 as shown in Fig. 5. The control message 510 indicates modulation parameters ^ and / or ^ employed by thetransmitter device 100 for the disclosed modulation scheme. Correspondingly, the receiver device 300 may be configured to receive the control message 510 from the transmitter device 300 and to derive modulation parameters ^ and / or ^ from thecontrol message 510. However, in other examples of the invention, the parameters ^ and / or ^ may be predetermined and henceknown to the receiver device 300 without the reception of the control message 510. In such cases the parameters ^ and / or ^may be standardized according to a communication standard such as 3GPP NR. When the modulation parameters ^ and / or ^is transmitted in control signaling the control message 510 may be in control information such as downlink control information(DCI) or uplink control information (UCI) depending on transmission direction.In BPSK, consecutive symbols may have a phase shift of either 0 or ^, while with QPSK consecutive symbols may have aphase shift of 0, ^ / 2, 3^ / 2, ^^ 3^ / 4. In conventional solutions, these constellations are known to suffer from zero crossings,i.e. a ^-phase change can happen from one modulation constellation symbol to the next one, resulting in a high PAPR. Therefore, ^ / 2-BPSK is often explained to provide lower PAPR by avoiding zero-crossing, i.e., consecutive symbol cannot have such ^-phase change as it is then systematically ^ / 2.This interpretation is correct for single carrier waveform but does apply to DFT-s-OFDM since DFT-s-OFDM can be expressedas a time-domain multiplexing scheme where consecutive symbols are transmitted by time-shifted sinc-like pulses. However,these DFT-s-OFDM pulses also have a phase difference such that consecutive pulses have a nearly ^ phase difference for timessamples between them where amplitude peak typically appears. Therefore, taking the phase effect of DFT-s-OFDM pulses,zero-crossing happens if the same constellation symbol is transmitted twice in a row, and ^ / 2-BPSK reduces the PAPR withDFT-s-OFDM not because it avoids two symbols having a ^ phase change, but actually because it avoids two symbols havinga 0-phase change. To reduce the PAPR, we therefore propose an adaptation of OQPSK modulation to DFT-s-OFDM that guarantees large phase difference among consecutive symbols. With the proposed modulation scheme for DFT-s-OFDM,consecutive symbols are selected to have phase shift to be either ^ / 2 or ^, and never 0.Specifically considering a 1 / ^ rate repetition code with ^ = 2 and offset ^ = 1 as discussed above, we define the followingnew modulation scheme as^^^^^For presentation clarity, we define indexing modulo ^^^ such that ^^^ = ^^^^ ^^, and ^^^ = ^^^^^^.Therefore, the ^th transmitted symbol can be written as The resulting modulation is illustrated in Table 4 below assuming ^^^ being even where we see that the stream of input bits issplit between the I-branch and Q-branch, then in each respective branch the bits are encoded with a 1 / ^ = 1 / 2 rate repetitioncode, and every repeated bit is flipped. Moreover, the repetition pattern between the I-branch and Q-branch is offset by ^ = 1bit. Finally, these two streams of encoded bits are jointly modulated as QPSK symbols. As it can be seen from Table 4, the bitflipping operation is equivalent to a sign change on the amplitude of the modulated I-branch or Q-branch and are thereforeequivalent. The bold lines in Table 4 and Table 5 illustrates the ^ = 2 occurrences of the bits due to repetition coding, andcorresponding repetition pattern on the I-branch and Q-branch of the QPSK symbols. The bold line also highlights the offsetbetween the repetition patterns of the I-branch and Q-branch.Table 4: Illustration of the new modulation scheme New modulation scheme Input Bits Constellation symbols QPSKI-branch Q-branch I-branch Q-branch^^ ^ ^^^^^^^^^^^^^ ^^ −^^^^^^(^^ , ^^ , … . , ^^^^^^)^ ^^^^ ^^ −^^ ^^^^ ^^^ ^^ −^^…. …. …. ….^^^^^^^^^^^^^^^^^^^^ ^^^^^^ − ^^^^^^^^^^^^^^^^^^^^ ^^^^^^ −^^^^^^ ^^^^^^Examples of the invention emulates the principle of OQPSK modulation with DFT-s-OFDM waveform: two consecutive QPSK symbols always share one repeated I or Q component, with an alternating sign change to account inherent DFT-s-OFDMwaveform property. Thus, the transmitter device 100 may flip at least one bit in each segment of ^ number of encoded bits ofthe encoded in-phase stream of bits 142 and flip at least one bit in each segment of ^ number of encoded bits of the encodedquadrature stream of bits 144. This can simply be generated at the transmitter device 100 with a bit repetition and flipping,which can be seen as a Manchester encoding of the bits, before QPSK modulation. The direct outcome is a low-PAPR signalin the range of 2~3dB.However, in conventional offset QPSK, the I-branch and Q-branch carries independent bit streams. The I-branch and Q-branchare offset by half of a symbol-period of symbol such that the phase shift of the signal is never larger than ^ / 2. This can be done in single carrier modulation or filter bank multicarrier systems, but not with DFT-s-OFDM. To circumvent this, here theoffset ^ is selected to be one symbol, and thus the data rate is reduced by half compared to QPSK. The bitflip operation that comes with repetition is used only to generate an alternative sign in the respective I or Q components of successive constellationsymbol, and thus to compensate the inherent nearly ^ phase change among consecutive DFT-s-OFDM pulses.Note that the combination of repetition and bit-flipping can be interpreted as Manchester coding, an example of implementationof the transmitter device 100 is thus Manchester coding before QPSK modulation.The new modulation scheme can also be seen as a form of differential QPSK encoding since the constellation symbol ^[^]depends on the previous symbol ^[^ − 1]. An illustration is provided in Fig.6 and 7 that show how transmitted QPSK symbolsare restricted depending on previous QPSK symbol.Fig. 6 shows that given the (^ − 1)th symbol on the stream of QSPK symbols, the next value of the ^th symbol is dependent ofthe previous symbol as well the parity of symbol index ^. Depending on if ^ is odd or even, the ^th symbol can only take valuesin a subset of two symbols, and never the value of the (^ − 1)th symbol.Fig. 7 shows an equivalent but different view of the differential aspect of the new modulation scheme. Here we see that thevalue of the (^ − 1)th bit on the I-branch and the Q-branch conditioned the value of the QPSK symbol to a restricted subset oftwo symbols. The subset depends on value of the (^ − 1)th bit and the parity of index ^, such that two cases of two subsets arepossible and alternating as a function of index ^. Finally, the transmitted QPSK symbol is one of the two symbols as a functionof the value of the ^th bit given the preconditioned subset according to the value of the (^ − 1)th bit. Nevertheless, as will beseen the symbol can be still decoded independently, removing the issue of error propagation from successive detection of differential encoders.After OFDM demodulation and equalization, the received symbols at the receiver device 300 are given as ^[^] = ^ ⋅ ^[^] +ICI[^] + ^[^] in Eq. (6).We consider ^^^even and will use a post-combining operation where each demodulated symbol is subtracted by the next,modulo ^^^, demodulated symbol as (^[^] − ^[^ + 1]).From this, it can be seen that the I and Q components can be recovered as Altogether, the soft-detection at the receiver device 300 can be written taking only the even ^ as This leads to an equivalent QPSK transmission which we can write as where and We see that that the combined symbols ^^can be decoded as QPSK symbols.While Eq. (30) shows the desired effect this interpretation is exact only for ZF equalizer for which I^CI ^ = 0. Otherwise, ingeneral, the post-combining interference term I^CI ^ is actually not purely interference and in fact contains also a desired signalcomponent. Considering also the desired signal component in I^CI^, the signal of Eq. (30) can be rewritten as where is an interference term As previously, treating the interference as noise and assuming the noise being Gaussian, the effective SNR is where the total signal power is and the noise power is independent of ^ and given as was the total signal power and noise power for QPSK and therefore ^^^ / 2 is the noise power for BPSK. In^^^^ = P^^^ and ^^ = ^, so that the effective SNR is ^^^^ the case of ZF, one can verify that Pgiven by SNR as expected. The above analysis is useful to perform soft-symbol de-mapping to be input into QPSK demodulation. According to Eq. (31),the channel gain is ^ ^ as given in Eq. (32), and the estimated noise power is P ^ ^^^^^ − ^ + ^[|n^^|^] where P^^^^ and ^[|n^^|^] are given in Eq. (32) and Eq. (35). These quantities can be used to compute loglikelihood ratio (LLR) values of the receivedbits. It can be remarked that the LLR computation would differ from conventional solutions by inserting a weigh function,^1 − cos^^ ^^^^^ in the averaging over subcarrier indices.Thus, in other words, demodulating the stream of in-phase symbols 342 and the stream of quadrature symbols 344 into thestream of soft QPSK symbols 350 may comprise compute LLRs of bits carried by the stream of in-phase symbols 342 and thestream of quadrature symbols 344, respectively, based on the formulawhere ^^^ is the number of subcarriers for transmitting the one or more DFT-s-OFDM symbols 160, and ^ is the subcarrierindex.A complete scheme of an example of the invention is summarized in Fig. 8 which shows a block diagram of a transmitterdevice 100 and a receiver device 300.The transmitter device 100 in Fig. 8 comprises an input block 170 configured to receive a stream of information bits fortransmission and a FEC block 172 configured to FEC encode the stream of information bits into a stream of bits 130. A splitting block 174 splits the stream of bits 130 into a stream of I-branch bits 132 and a stream of Q-branch bits 134. An encoder andmodulation block 176 firstly encodes the stream of I-branch bits 132 and a stream of Q-branch bits 134 into encoded I-branchstream of bits 142 and encoded Q-branch stream of bits 144 and thereafter jointly modulates the encoded I-branch stream ofbits 142 and encoded Q-branch stream of bits 144 into a stream of QPSK symbols 150 in the QPSK modulator block. Thestream of QPSK symbols 150 is converted from serial-to-parallel (S / P) and inputted into a DFT-s-OFDM modulation block 178. Modulating the stream of QPSK symbols into one or more DFT-s-OFDM may comprise in examples of the invention: DFTmodulating the stream of QPSK symbols 150 into a stream of Fourier coefficients 152 in a DFT block; FDSS the stream ofFourier coefficients 152 into a stream of shaped Fourier coefficients 154 based on a bell-shaped FDSS window in a FDSSblock; and IFFT modulating the stream of shaped Fourier coefficients 154 into the one or more DFT-s-OFDM symbols 160 ina IFFT block. Finally, the one or more DFT-s-OFDM symbols 160 are transmitted to a receiver device 300 over a radio channel.Before transmission, a CP may be added to each DFT-s-OFDM symbol 160 in order to combat time-dispersive channel.The receiver device 300 in Fig. 8 comprises a DFT-s-OFDM demodulation block 370 configured to receive the one or moreDFT-s-OFDM symbols 160 transmitted by the transmitting device 100. The reverse operations are performed by the DFT-s-OFDM demodulation block 370 in the receiver device 300 compared to what is performed by the DFT-s-OFDM modulationblock 178 in the transmitter device 100 which results in that a stream of demodulated constellation symbols 330 is outputtedfrom the DFT-s-OFDM demodulation block 370 into a combining and splitter block 372 via a FFT block, an equalizer blockand a IDFT block.In the combining and splitter block 372 the ^ number of demodulated constellation symbols 330 are successively combinedinto a stream of combined symbols 332 and thereafter a first half of the stream of combined symbols 332 are split into a stream of in-phase symbols 342 (I-branch) and a second half of the stream of combined symbols 332 into a stream of quadraturesymbols 344 (Q-branch). The successive combining of the ^ number of demodulated constellation symbols 330 into the streamof combined symbols 332 may in examples of the invention comprise averaging the ^ number of demodulated constellationsymbols 330 with an alternating sign change. Since the transmitter has encoded the bits with 1 / ^ repetition coding, the receiverdevice 300 can detect each bit from ^ QPSK symbols. By combining ^ symbols carrying the same bit enables to improve theSNR from diversity gain. Several combining methods exist that may be used, e.g., selection combining, equal gain combining,or maximum ratio combining. As with DFT-s-OFDM, the demodulated constellation symbols are all attenuated by the samechannel gain, maximum ratio combining reduces to a simple averaging, except that before averaging, the bit flipping operationof equivalent alternating sign change made at the transmitter device 100 needs to be compensated at the receiver device 300.Moreover, the splitting of the first half of the stream of combined symbols 332 into the stream of in-phase symbols 342 and the second half of the stream of combined symbols 332 into the stream of quadrature symbols 344 may be based on the offsetof ^ number of combined symbols. This is needed to compensate for the offset made at the transmitter device 100 and forreusing existing QPSK demodulation for decoding the transmitted bits.Different equivalent receiver device implementations are possible. For example, the real and imaginary part of the demodulatedconstellation symbols 330 could first be split before combining. Then combining will be applied to the I- and Q-branch,respectively.The stream of in-phase symbols 342 and the stream of quadrature symbols 344 are inputted into a demodulation and decoderblock 376 in which the stream of in-phase symbols 342 and the stream of quadrature symbols 344 are jointly demodulated intoa stream of soft QPSK symbols 350. The demodulating of the stream of in-phase symbols 342 and the stream of quadraturesymbols 344 into the stream of soft QPSK symbols 350 is typically done by computing LLRs of bits carried by the stream ofin-phase symbols 342 and the stream of quadrature symbols 344, respectively, Finally, the stream of soft QPSK symbols 350 is decoded into a stream of decoded bits in a decoding block 378. FDSS is a data-independent technique to reduce the PAPR of single carrier waveforms. In NR, ^ / 2-BSPK was introduced forDFT-s-OFDM with looser spectral flatness requirement in order to enable the usage of a FDSS window to further reduce thePAPR. FDSS is a transparent method in the sense that no explicit FDSS window is specified in 3GPP standard. The PAPR reduction with FDSS comes at the cost of a SNR loss. The SNR loss increases with coding rate, which makes FDSS more effective for low coding rates. We will use FDSS windows as considered by different companies in 3GPP, which are theruncated root-raised-cosine (RRC) with parameters (0.5, -0.65), and 3-tap filters constructed as ^ ^ =^^^^t [ ] ∑^^^^ ^[^]^^^^^^,0 ≤ ^ ≤ ^^^, where ^ = [−0.28, 1, −0.28] or [−0.335, 1, −0.335]. For convenience, we will label them as ‘RRC’, ‘3tap1’,and ‘3tap2’, respectively. In order to keep transmit power unchanged the window coefficients should be normalized in orderto satisfy ∑^^^^^^^^|^[^]|^ = ^^^ .The Cauchy-Swartz inequality indicates that using a FDSS (ignoring the normalization factor) that would match ^^[^] wouldincrease the desired signal power according to Eq. (32), and thus at the same time mitigate the interference. It is observed that this indeed provides the best error rate performance with MMSE i.e., performance without power de-rating from envelope fluctuation. Meanwhile, according to current 3GPP standard such FDSS window could not be used as it would not fulfill spectrum mask.Therefore, in examples of the invention, the modulating of the stream of QPSK symbols 150 into the one or more DFT-s-OFDM symbols 160 comprises: DFT modulate the stream of QPSK symbols 150 into a stream of Fourier coefficients 152;FDSS the stream of Fourier coefficients 152 into a stream of shaped Fourier coefficients 154 based on a bell-shaped FDSSwindow; and finally to IFFT modulating the stream of shaped Fourier coefficients 154 into the one or more DFT-s-OFDMsymbols 160. Thus, compared to conventional solutions, all terms in the SNR expression of Eq. 33 are shaped before averaging oversubcarrier indices according to the same FDSS window function which is based on: ^^[^] = ^1 −^^ ^^, where ^^^is the number of subcarriers used for transmitting the one or more DFT-s-OFDM symbols 160, and ^ is the subcarrier index. This window is a bell-shape window with a maximum at the middle subcarrier, similarly as FDSS window for PAPR reductionand can be interpreted as a linear function of or as a transformation of 1 − cos Thus, the bell-shaped FDSS window is based on where ^^^is the number of subcarriers used for transmitting the one or more DFT-s-OFDM symbols 160, and ^ is the subcarrierindex.This may imply that the stream of Fourier coefficients 152 are cyclically shifted by an amount ^^^ / 2. A more formalizedexpression for the bell-shape window may be defined as such that the maximum of the filter is ^[^^^ / 2] = 1 and the maximum power attenuation is ^[0] = ^. Taking ^ = 1corresponds to no FDSS, while ^ = 0 match ^^[^] after normalization.With the current 3GPP spectrum mask, the maximum tolerable attenuation is 14 dB which correspond to ^ = 10^^.^ ≈ 0.1995.As shown in the Appendix, it appears that this almost exactly matches the 3tap2 filters previously used in conventionalsolutions. We observed for MMSE that ^^[^] with ^ = 10^^.^, and thus the 3tap2 window, performs very close to the performance with ^^[^]. If considering ZF, the FDSS window changes only the noise power Eq. (35). In this case, we observedthat the performance is improved by decreasing ^ starting from ^ = 1 until ^ ≈ 0.5, then the performance degrades from thispoint. This indicates a way to minimize the error performance, the resulting PAPR / CM may however not be the best. Therefore, if taking also into account the power rating, selecting the best FDSS window may be a trade-off between SNR gain in error rates and PAPR / CM reduction which may change depending on the considered transmission configurations.In summary, the new modulation scheme can be decoded simply by a pairwise-symbol combiner, with precisely ^^^ extraadditions, before QPSK demodulation. A surprising benefit from this combining is that by using a FDSS window at thetransmitter device 300 not only the PAPR is improved, but also provides a SNR gain in the BLER performance. This is theopposite to conventional solutions for which a FDSS window always brings PAPR reduction at the cost of some SNR loss in the BLER performance. The observed SNR gain of the new modulation scheme combined with FDSS is in the range of 0.5~1.7 dB, where it gets larger with larger transmission bandwidth, and channel with little channel frequency selectivity as often considered for NTN.An even number of subcarriers provides benefit based from the exemplary solution while being practically more relevant. Thus,the transmitter device 100 is in examples of the invention configured to transmit the one or more DFT-s-OFDM symbols 160on an even number of subcarriers. Correspondingly, the receiver device 300 may be configured to receive the one or moreDFT-s-OFDM symbols 160 on the even number of subcarriers. However, when considering an odd number of subcarriers, there are two bits, one each on the I-branch and Q-branch, which cannot be repeated. Then the PAPR / CM would not be as much reduced, but the PAPR / CM reduction becomes neverthelesssimilar as the bandwidth increases. For example, it is observed that by removing one subcarrier from a 24-subcarrier allocationthe PAPR increases by 1dB and the CM by 0.5 dB. If removing one subcarrier from a 288-subcarrier allocation, the PAPR thenincreases by about half a dB and the CM by a small fraction of dB. Also, with an odd number of subcarriers, the imaginary part of the first received symbol and real part of the last symbol form an independent QPSK symbol that cannot be combined withany other I and Q components from other subcarriers. Therefore, the effective SNR of this specific QPSK symbol is as givenfor conventional solutions, different to the effective SNR for other symbols derived above, and must be demodulatedspecifically as such. This would have a marginal effect, that would again disappear as the subcarrier allocation is increased.Alternatively, in this case, one could employ repetition coding for the last 2 bits with rate 1 / 3, i.e., ^ = 3 instead of ^ = 2,PAPR would be low, but then there is a reduced rate where for two bits a stronger SNR can be obtained following a 3-symbolcombining.In the disclosure, the bit-flip / sign-change is applied systematically on the repeated bit. Alternatively, other examples may beobtained by applying the bit flip on the original bit instead of the repeated bit. This bit flip can also be applied differently onthe I-branch and Q-branch. Any solution where for the I-branch or Q-branch the bits are repeated with a bit flip would providethe same performance. For example, an alternative example is given as The ^th transmitted symbol can be written in a compact form as and the resulting transmission is illustrated in Table 5 below where we see that the input bit stream is split between the I-branchand Q-branch, and in each respective branch the bits are encoded with a 1 / 2 rate repetition code, but compared to Table 4 thebit flipping operations in Table 5 are aligned between the I-branch and Q-branch such that for the I-branch it is the repeated bitwhich is flipped while for the Q-branch it is the original bit which is flipped. Thus, at least one bit in each segment of ^ numberof encoded bits of the encoded in-phase stream of bits 142 is flipped and at least one bit in each segment of ^ number ofencoded bits of the encoded quadrature stream of bits 144 is flipped. As in Table 4 the repetition pattern between the I-branchand Q-branch in Table 5 is offset by ^ = 1 bit, and these two streams of encoded bits are jointly modulated as QPSK symbols.Again, the bit flipping operation is equivalent to a sign change on the amplitude of the modulated I-branch or Q-branch, andthus the sign change is alternating over the stream of QPSK symbols.Table 5: Illustration of another example of the new modulation schemeNew modulation scheme Input Bits Constellation symbolsI-branch Q-branch I-branch Q-branch^^^^^^^^^^^^^^^^(^^ , ^^ , … . , ^^^^^^)^ ^^^^ ^^^ −^^ −^^^^^^^^^^…. …. …. ….^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ ^ ^^^^^^^^^^^^^ −^^^^^^ −^^^^^^An advantage of this example of the invention is that it could be implemented by only repeating the bits and applying analternating + / -1 mask sequence on the QPSK symbols before DFT precoding. This is equivalent to cyclic shift the stream ofFourier coefficients 152 by ^^^ / 2. Thus, the transmitter device 100 may alternate a sign of at least one QPSK symbol in eachsegment of ^ number of QPSK symbols of the stream of QPSK symbols 150. The corresponding receiver device 300 combiner is given by first removing the sign change as ^[̃^] = (−1)^ ^[^]. (38)and then averaging consecutive symbols as In 3GPP TS 38.214 “NR; Physical layer procedures for data,” version 17.6.0, Release 17 there are two MCS tables for uplinkdata with DFT-s-OFDM. In each, there are 32 MCS indices of which 4 MCS indices (one for each modulation order) arereserved and used for re-transmission. The lowest MSC indexes can be selected to be either ^ / 2-BPSK or QPSK based on aconfiguration parameter. For ^ / 2-BPSK, the spectral efficiency of the MCS is matching the coding rate, while if QPSK isselected, the coding rate 1 / ^ is halved compared to ^ / 2-BPSK to get the same spectral efficiency. In the high spectralefficiency ^ / 2-BPSK can only be used for the two first MCSs with spectral efficiency 0.23 and 0.30 bits per channel use(bpcu), while QPSK is used also for the next 8 MSC indices, but only 3 of them are for a spectral efficiency above one. In thelow spectral efficiency ^ / 2-BPSK can be used in six MSCs, with spectral efficiency from 0.05 to 0.19 bpcu. QPSK is used in the next 10 MCS with spectral efficiency from 0.24 to 1.31 bpcu (so with coding rate from 0.12 to 0.66), in which again only three of them are for a spectral efficiency above one. In total, 13 of the unreserved 28 MCS indices, i.e., almost half of the MCS indices, are for supporting a spectral efficiency below one bpcu. PAPR is an indication of the envelope fluctuation of a signal. Peaks in signal envelope either cause signal distortion by reachingthe non-linear region of the power amplifier; or enforce the need for power de-rating to avoid such distortion which directlydecreases the link budget.3GPP NR specifies maximum power reduction (MPR) values that limits the power de-rating (also referred as output back-off,offset, reduction) that can be used to fulfill conformance tests which includes several RF requirements that quantifies signaldistortion and adjacent channel interference. Necessary power de-rating from different power-efficient techniques is evaluated in 3GPP NR by RAN4 working group according to RF simulations where one decreases iteratively the transmit power from the maximum power until all requirements are satisfied. The power de-rating depends on the bandwidth allocation size and position within the channel bandwidth, typically being higher for large bands or small bands allocated at the channel band edge. Thus, in the corresponding NR specification, edge allocations have larger MPR values compared to inner allocations, and larger allocated bandwidth have additional MPR margins. In general, the higher is the modulation order, the higher is the PAPR, and thus the higher is the allowed MPR. The cubic metric (CM) was introduced in early LTE standardization based on non-linear factor of power amplifier modellingand argued that after linear regression it predicts more accurately the power de-rating than PAPR. Since then, 3GPP RAN1 hasbeen using the same CM fitting on old measurements as a proxy to compare power de-rating from different schemes, but PAPR remains always discussed. Neither CM nor PAPR completely predict realistic power de-rating values. For example, both CM and PAPR of QPSK with a given FDSS are worse than with clipping, while the simulated power de-rating is the opposite. Power de-rating as a function of the bandwidth allocation and size are also not captured neither by the CM and PAPR. For thisreason, in recent Rel.-18 work item on coverage enhancement, power de-rating estimation based on CM was agreed at first tobe used by RAN1 before RAN4 evaluation, but major companies pointed out the limitation of CM / PAPR and provided directly RF simulations in RAN1 instead. In the disclosure, the power de-rating of each transmission scheme will be estimated according to its respective 99-percentile CM values as similarly done by RAN1 at the beginning of Rel.-18 coverage enhancement work item. However, we stress the following issue in CM-based modeling indicating that CM underestimate power de-rating gains for QPSK-FDSS and likely for the same reason to the disclosure. PAPR and CM are not always consistent metrics, and CM seems to suffer from a specific misfunctioning behavior for QPSKwith FDSS. For ^ / 2-BPSK, both PAPR and CM always decreases with FDSS, consistently with power de-rating simulation:the more bended is the FDSS window the lower gets the PAPR / CM and power de-rating. However, for QPSK, there is an inconsistent behavior between PAPR and CM with FDSS, and notably the CM can get worse with FDSS contradicting the PAPR performance and realistic power de-rating simulation. Indeed, it can be verified that CMfor QPSK slightly decreases at first with some shaping but increases again with more shaping. This is inconsistent not onlywith PAPR but also power der-rating simulation that shows a decreasing trend with FDSS. The CM difference between QPSK and ^ / 2-BPSK is about 1.5dB which matches very well the simulated power de-rating gain. However, therein the CM gain of FDSS for QPSK with RRC window is only 0.1dB while the simulated power de-rating gains are much larger between 0.3 to 0.8 dB. If using spectrum extension for QPSK, PAPR and CM become more consistent but the CM still underestimates the power de-rating gain of FDSS-SE for QPSK: 0.9 dB compared to 1.2~1.6 dB. Similarly, clear power de-rating gain are also shown for QPSK with several FDSS (without spectrum extension), while the corresponding CM gains are small or zero. A power de-rating gain of 0.3~0.9 dB for QPSK-FDSS with RRC window is shown, and for the more bended 3tap1 filter, the power de-rating gain gets even larger, up to 1.2 dB, while the CM gain is reduced to zero. An effect of examples of the invention is to improve the error performance compared to conventional solutions with and without power de-rating. For comparison we will take the baseline LOS channel model NTN-TDL-C with delay scaling of 3ns which is the mean delay spread in LOS rural scenario at S-band with 30° elevation angle, which corresponds to the beam edge of a LEO satellite. This channel has little frequency selectivity, and we assume block fading and perfect channel estimation. The S- band has a system bandwidth of 30 MHz which could support 166 resource blocks (RB) of 12 subcarriers. Recently, 3GPPconsidered for uplink coverage evaluation of handled UE very narrow band, 2RB. UEs with stronger SNR condition or / andwith higher power class such as VSAT terminals have been considered transmitting on larger bands up the channel bandwidth.For comparison we will consider two equivalent allocations of 288 resource elements: a narrow band allocation with 2 RB and12 OFDM symbols (OS), and a wideband allocation with 24 RBs and 1 OFDM symbol. All simulations are with MMSE, except if specified.Fig. 9 shows the PAPR for 2 RBs and 12 OS allocation where the x-axis shows PAPR in dB and the y-axis showscomplementary cumulative distribution function (CCDF). The new modulation scheme, denoted “RO-QPSK” (repeated andoffset QPSK), without FDSS leads to a PAPR of 3.5 dB, and with FDSS to a PAPR of 1.7 dB using RCC FDSS window. With^ / 2-BPSK, the PAPR is much higher without FDSS: 5.5 dB, but similar with FDSS: ~2 dB with 3tap2. If looking at the CM,the comparison is inconsistent with PAPR. Notably, the CM of the new modulation scheme is increased if using an FDSS window, up to 0.5 dB, which is the opposite behavior than for PAPR. For ^ / 2-BPSK, comparison between PAPR and CM areconsistent. An important observation is that both the PAPR and the CM of OQSPK gets better with the larger bandwidth. Forexample, considering a 24 RB allocation, the PAPR of the new modulation scheme decreases and is about 2 dB without FDSS and less using 3tap1 window. This may be a consequence that the phase difference between consecutive DFT-s-OFDM pulses get closer to π as ^^^→ ∞.Fig. 10 shows the maximum achievable spectral efficiency of the different considered constellations based on average mutualinformation where the x-axis shows SNR in dB and the y-axis the spectral efficiency in bpcu. The mutual information issimulated according to the derived transmission equations assuming Gaussian interference, averaged over multiple channel realizations. Fig.10 is obtained with 2 RBs and NTN-TDL-C channel. The new modulation scheme, denoted “RO-QPSK”, can provide almost the spectral efficiency than ^ / 2-BPSK with a negligible difference. If considering FDSS, here with 3tap2window, we see that the spectral efficiency of ^ / 2-BPSK is decreased with a gap that increase with SNR. Contrary to the newmodulation scheme, the spectral efficiency improves if using the FDSS window, such that it can also outperform ^ / 2-BPSKand QPSK up to a spectral efficiency of 0.45 bpcu. The maximum SNR gain over conventional solutions that can be observed1.7dB.Fig. 11 shows coded block error rate (BLER) simulation with 2 RB allocations, for 0.05 bpcu spectral efficiency, and NTN-TDLC LOS rural. The maximum spectral efficiency from the mutual information in Fig. 10 can only be approached withinfinite codelength. Fig. 11 shows error rate performance with finite block length. The channel code is LDPC with coding ratefor ^ / 2-BPSK and the new modulation scheme corresponding to the desired spectral efficiency, and half-rate QPSK for fair comparison. The target BLER is 10%, for smaller BLER the gains may be reduced but we observed that in such casesimprovement can be obtained with MF equalizer. In Fig. 11, the left plot is conventional BLER performance as a function ofSNR, while the right plot is a BLER performance as a function of SNR with power de-rating where the power de-rating of eachscheme is according to its respective 99-percentile CM values. As explained before, such power rating estimation via CM is only indicative in general and may not reflect even relative RF simulation comparison. For the new modulation scheme with FDSS, a similar phenomenon happens as for QPSK with FDSS, the PAPR and CM have contradicting behavior and thus by using CM the gains with power de-rating gains are likely pessimistic.The SNR gain between the best performance of the new modulation scheme compared to the best performance amongconventional solutions are given in Table 6, considering also 24 RB allocations and NTN-TDL-A channel. The consideredspectral efficiencies are 0.05, 0.1 and 0.3, as supported for MCS compatible with ^ / 2-BPSK in current NR standards. WithoutFDSS, the new modulation scheme has a similar BLER than ^ / 2-BPSK and QPSK for 0.05 and 0.1 spectral efficiencies, buta noticeable dB loss of 0.6dB SNR at 1% BLER for 0.3 bpcu with 2 RB allocation. This SNR loss is however again negligible for 0.3 bpcu with the larger 24 RB allocation. With FDSS, the new modulation scheme can outperform all conventionalsolutions even the one without FDSS. Using the new modulation scheme with FDSS can provide SNR gain in the range 1.1~1.7dB at 1% BLER over conventional solutions, consistent with the mutual information analysis above. The gain is larger withthe lowest spectral efficiency and larger allocation, and the smallest for 0.3 bpcu with 2 RB allocation. If considering power de-rating based on CM performance, the gain of the new modulation scheme with FDSS compared to conventional solutions is smaller in the range of 0.6~1.3 dB as the CM increases from the new modulation scheme with FDSS which may not berepresentative as on the contrary the PAPR improves. Still, the new modulation scheme always provides the best performance.We also considered performance with the NTN-TDL-A channel with delay scaling of 100 ns which is a NLOS channel model and a thus a more frequency selective channel. Note that a 100 ns delay scaling is a large delay spread for satellite modelling, corresponding for example to the mean delay spread at 30° elevation angle (cell edge) in NLOS dense urban scenario in S- band. With such channel assumption, a larger SNR range is needed to support the same spectral efficiency, which further limit the data rate that could be supported for the same SNR link. With highly frequency selective channel, the gain of the disclosedmethod is reduced, notably for the largest bandwidth allocation where the effect of frequency selectivity is more pronounced.The SNR gain range is 0.7~1.4 dB and 0.6~0.9 dB with and without power rating. Again, the new modulation scheme alwaysperforms best. The SNR gains in BLER from the new modulation scheme are observed above for narrow band and / or channelwith little frequency selectivity, following current NTN simulation scenarios.In summary as shown on Table 6, for narrow band as in NTN UL, the new modulation scheme would have an SNR gain overQPSK and ^ / 2-BPSK, in addition to a potential additional power de-rating gain over QPSK depending on the configuration. Using CM as a power de-rating estimation, which we stressed to be pessimistic for the disclosed solution as the CM increaseswith FDSS contrary to the PAPR, the new modulation scheme still systematically outperforms conventional solutions with SNR gains of 0.6~1.3 dB. So, in general, the disclosed solution is shown to provide overall the best performance with andwithout power de-rating, and thus to improve the link budget for both cell-edge and non-cell edge UEs for low spectralefficiencies as supported by ^ / 2-BPSK in NR.Table 6SNR gain [dB] SNR gain [dB] Scenario SE@1% BLER @1% BLER with power de- [bpcu] rating NTN-TDLC LOS rural – 2RB / 14OS 0.05 1.4 0.7NTN-TDLC LOS rural – 2RB / 14OS 0.1 1.3 0.7NTN-TDLC LOS rural – 2RB / 14OS 0.3 1.1 0.6NTN-TDLC LOS rural – 24RB / 1OS 0.05 1.6 1.0NTN-TDLC LOS rural – 24RB / 1OS 0.1 1.7 1.0NTN-TDLC LOS rural – 24RB / 1OS 0.3 1.4 1.3NTN-TDLA NLOS Dense Urban – 2RB / 14OS 0.05 1.4 0.7NTN-TDLA NLOS Dense Urban – 2RB / 14OS 0.1 1.3 0.7NTN-TDLA NLOS Dense Urban – 24RB / 1OS 0.05 0.5 0.6NTN-TDLA NLOS Dense Urban – 24RB / 1OS 0.1 0.7 0.9Furthermore, any method according to examples of the invention may be implemented in a computer program, having code means, which when run by processing means causes the processing means to execute the steps of the method. The computer program is included in a computer readable medium of a computer program product. The computer readable medium may comprise essentially any memory, such as previously mentioned a ROM, a PROM, an EPROM, a flash memory, an EEPROM, or a hard disk drive.Moreover, it should be realized that the transmitter device 100 and the receiver device 300 comprise the necessarycommunication capabilities in the form of e.g., functions, means, units, elements, etc., for performing or implementing examples of the invention. Examples of other such means, units, elements and functions are: processors, memory, buffers, control logic, encoders, decoders, rate matchers, de-rate matchers, mapping units, multipliers, decision units, selecting units, switches, interleavers, de-interleavers, modulators, demodulators, inputs, outputs, antennas, amplifiers, receiver units, transmitter units, DSPs, TCM encoder, TCM decoder, power supply units, power feeders, communication interfaces, communication protocols, etc. which are suitably arranged together for performing the solution.Therefore, the processor(s) of the transmitter device 100 and the receiver device 300 may comprise, e.g., one or more instancesof a CPU, a processing unit, a processing circuit, a processor, an ASIC, a microprocessor, or other processing logic that may interpret and execute instructions. The expression “processor” may thus represent a processing circuitry comprising a plurality of processing circuits, such as e.g., any, some or all of the ones mentioned above. The processing circuitry may further perform data processing functions for inputting, outputting, and processing of data comprising data buffering and device control functions, such as call processing control, user interface control, or the like. Finally, it should be understood that the invention is not limited to the examples described above, but also relates to andincorporates all examples within the scope of the appended independent claims.AppendixConsidering a 3-tap filter with ^ = [−^, 1, −^] such thatThe phase is irrelevant and can be ignored. When comparing to spectrum mask the filter should be scaled such that is maximumis 1, which leads here to

Claims

CLAIMS 1. A transmitter device (100) configured to: split a stream of bits (130) into an in-phase stream of bits (132) and a quadrature stream of bits (134); encode the in-phase stream of bits (132) and the quadrature stream of bits (134) with a 1 / ^ rate repetition code into anencoded in-phase stream of bits (142) and an encoded quadrature stream of bits (144), wherein the encoded quadrature streamof bits (144) is offset in relation to the encoded in-phase stream of bits (142) with ^ number of bits, or vice versa, where ^ and^ are positive integers;jointly modulate the encoded in-phase stream of bits (142) and the encoded quadrature stream of bits (144) into a stream of Quadrature Phase Shift Keying, QPSK, symbols (150); modulate the stream of QPSK symbols (150) into one or more Discrete Fourier Transform spread Orthogonal Frequency Division Multiplexing, DFT-s-OFDM, symbols (160); and transmit the one or more DFT-s-OFDM symbols (160).

2. The transmitter device (100) according to claim 1, wherein ^ is dependent on ^.

3. The transmitter device (100) according to claim 1 or 2, wherein 1 ≤ ^ ≤ ^ − 1 modulo ^.

4. The transmitter device (100) according to claim 3, wherein ^ = 2.

5. The transmitter device (100) according to any one of the preceding claims, configured to: flip at least one bit in each segment of ^ number of encoded bits of the encoded in-phase stream of bits (142) and flipat least one bit in each segment of ^ number of encoded bits of the encoded quadrature stream of bits (144); oralternate a sign of at least one QPSK symbol in each segment of ^ number of QPSK symbols of the stream of QPSKsymbols (150).

6. The transmitter device (100) according to any one of the preceding claims, wherein the modulating of the stream of QPSKsymbols (150) into the one or more DFT-s-OFDM symbols (160) comprises:DFT modulate the stream of QPSK symbols (150) into a stream of Fourier coefficients (152);Frequency Domain Spectral Shaping, FDSS, the stream of Fourier coefficients (152) into a stream of shaped Fouriercoefficients (154) based on a bell-shaped FDSS window; andInverse Fast Fourier Transform, IFFT, modulating the stream of shaped Fourier coefficients (154) into the one or moreDFT-s-OFDM symbols (160).

8. The transmitter device (100) according to claim 7, configured to:cyclic shift the stream of Fourier coefficients (152) by ^^^ / 2.

9. The transmitter device (100) according to any one of the preceding claims, configured to: transmit the one or more DFT-s-OFDM symbols (160) on an even number of subcarriers.

10. The transmitter device (100) according to any one of the preceding claims, wherein the stream of bits (130) is a stream of information bits (128) encoded with a Forward Error Correction, FEC, code.

11. The transmitter device (100) according to any one of the preceding claims, configured to:transmit a control message (510) to a receiver device (300), the control message (510) indicating ^ and / or ^.

12. A receiver device (300) configured to: receive one or more DFT-s-OFDM symbols (160); demodulate the one or more DFT-s-OFDM symbols (160) into a stream of demodulated constellation symbols (330); successive combining ^ number of demodulated constellation symbols (330) into a stream of combined symbols (332),where ^ is a positive integer;split a first half of the stream of combined symbols (332) into a stream of in-phase symbols (342) and a second half of the stream of combined symbols (332) into a stream of quadrature symbols (344); jointly demodulate the stream of in-phase symbols (342) and the stream of quadrature symbols (344) into a stream ofsoft QPSK symbols (350); anddecode the stream of soft QPSK symbols (350) into a stream of decoded bits (352).

13. The receiver device (300) according to claim 12, wherein the successive combining of the ^ number of demodulatedconstellation symbols (330) into the stream of combined symbols (332) comprises: average the ^ number of demodulated constellation symbols (330) with an alternating sign change.

14. The receiver device (300) according to claim 12 or 13, wherein the splitting of the first half of the stream of combined symbols (332) into the stream of in-phase symbols (342) and the second half of the stream of combined symbols (332) into thestream of quadrature symbols (344) is based on an offset of ^ number of combined symbols, where ^ is positive integer.

15. The receiver device (300) according to any one of claims 12 to 14, wherein the demodulating of the stream of in-phasesymbols (342) and the stream of quadrature symbols (344) into the stream of soft QPSK symbols (350) comprises:compute loglikelihood ratios, LLRs, of bits carried by the stream of in-phase symbols (342) and the stream of quadrature symbols (344), respectively, based on:where ^^^ is the number of subcarriers for transmitting the one or more DFT-s-OFDM symbols (160), and ^ is the subcarrierindex.

16. The receiver device (300) according to any one of claims 12 to 15, configured to:receive the one or more DFT-s-OFDM symbols (160) on an even number of subcarriers.

17. The receiver device (300) according to any one of claims 12 to 16, configured to:receive a control message (510) from a transmitter device (300), the control message (510) indicating ^ and / or ^.

18. The receiver device (300) according to any one of claims 12 to 16, wherein ^ and / or ^ are predetermined.

19. A method (200) for a transmitter device (100), the method (200) comprising: splitting (202) a stream of bits (130) into an in-phase stream of bits (132) and a quadrature stream of bits (134);encoding (204) the in-phase stream of bits (132) and the quadrature stream of bits (134) with a 1 / ^ rate repetition codeinto an encoded in-phase stream of bits (142) and an encoded quadrature stream of bits (144), wherein the encoded quadraturestream of bits (144) is offset in relation to the encoded in-phase stream of bits (142) with ^ number of bits, or vice versa, where^ and ^ are positive integers;jointly modulating (206) the encoded in-phase stream of bits (142) and the encoded quadrature stream of bits (144) intoa stream of QPSK symbols (150); modulating (208) the stream of QPSK symbols (150) into one or more DFT-s-OFDM symbols (160); and transmitting (210) the one or more DFT-s-OFDM symbols (160).

20. A method (400) for a receiver device (300), the method (400) comprising: receiving (402) one or more DFT-s-OFDM symbols (160); demodulating (404) the one or more DFT-s-OFDM symbols (160) into a stream of demodulated constellation symbols (330); successive combining (406) ^ number of demodulated constellation symbols (330) into a stream of combined symbols(332), where ^ is a positive integer; splitting (408) a first half of the stream of combined symbols (332) into a stream of in-phase symbols (342) and a second half of the stream of combined symbols (332) into a stream of quadrature symbols (344); jointly demodulating (410) the stream of in-phase symbols (342) and the stream of quadrature symbols (344) into a stream of soft QPSK symbols (350); and decoding (412) the stream of soft QPSK symbols (350) into a stream of decoded bits (352).

21. A computer program with a program code for performing a method according to claim 19 or 20 when the computer programruns on a computer.

Images