Multiple-input, multiple-output (MIMO) techniques usually involve some kind of pre-processing at the transmitter, namely when singular value decomposition (SVD) techniques are employed. However, the precoding operations in general and the ones associated to SVD techniques in particular can substantially increase the envelope fluctuations of the transmitted signal, which leads to an increased sensitivity to transmitter nonlinearities and inherent performance degradation. However, recent research demonstrates that nonlinear distortion can be seen as useful information, provided that an optimum detection is considered. In this work, we present an analytical approach to obtain the asymptotic optimum performance of MIMO-SVD systems with strong nonlinear distortion effects1.

The need for higher data rates in wireless communications has resulted in a significant effort in what regards to the design of transmission techniques and waveforms. This design can be very challenging thanks to the strong trade-off between the spectral efficiency and the energy efficiency, which comes from the fact that spectral-efficient waveforms usually presents a very large peak-to-average power ratio (PAPR), requiring a linear amplification for preserving the linearity of the transmitted signals. Nevertheless, obtaining spectral- and energy-efficient wireless communications will undoubtedly be one of the main challenges of both 5G and beyond 5G (B5G) systems. In this work, we propose a new amplification scheme named quantized digital amplification (QDA). The QDA is based on the decomposition of a high-PAPR signal into constant-envelope components and allows to obtain very high energy efficiencies, being suitable for spectrally-efficient transmission techniques such as orthogonal frequency-division multiplexing (OFDM). By presenting a set of results of the QDA’s prototype, we demonstrate that it is possible to obtain amplification efficiencies way above the ones of the state-of-the art amplification schemes, while preserving the linearity of the transmitted signals.1

Besides traditional orthogonal multiple access (OMA), 5G and other future wireless communication systems will also involve non-orthogonal multiple access (NOMA). In that context, the combination of NOMA systems with the popular orthogonal frequency-division multiplexing (OFDM) will arise naturally in highly spectral efficient communications. Since OFDM signals are very prone to nonlinear distortion effects, the receiver design of NOMA-OFDM systems should be modified accordingly. In this work, we consider an efficient, nonlinear power amplification for NOMA-OFDM systems and we propose different reception strategies to cope with the nonlinear distortion. It is shown that the performance of conventional power-domain NOMA is very poor under nonlinear transmission scenarios. However, it is also demonstrated that Bussgang receivers combined with optimum detection schemes might be an excellent alternative1.

Non-orthogonal multiple access (NOMA) is being considered a key technology for 5G systems. In this work, we consider power-domain NOMA schemes for downlink orthogonal frequency-division multiplexing (OFDM) transmissions. Moreover, we consider the use of an efficient, nonlinear power amplification. We analyze the nonlinear distortion effects associated with the power amplifier and show that the nonlinear distortion can be severe, preventing the use of conventional NOMA receivers. In that context, we propose a receiver that estimates and cancels the nonlinear distortion and enables the implementation of power-domain NOMA schemes in strongly nonlinear OFDM scenarios.

Massive multiple-input multiple-output (mMIMO) systems are very important for 5G since they are one of the main enablers of the targeted huge capacity gains. On the other hand, the power consumption at the mobile terminals should be as low as possible, making single-carrier waveforms, such as single carrier with frequency-domain equalization (SC-FDE), particularly interesting, since the transmit signals can have a low peak-to-average power ratio (PAPR), allowing a highly efficient power amplification. However, the combination of mMIMO systems with SC-FDE modulations gives rise to implementation difficulties at the receiver side. On one hand, the equalization procedure should avoid matrix inversions, but this might lead to performance degradation. On the other hand, low-resolution analog-to-digital converters (ADCs) should be employed in each receive branch to reduce the implementation complexity, which might lead to severe nonlinear distortion effects. In this work, we study analytically the performance of mMIMO systems based on SC-FDE schemes employing low-complexity, iterative FDE receivers, and low-resolution ADCs. It is shown that the performance degradation associated to low-resolution ADCs can be tolerable if the number of receive antennas is larger than the number of transmit antennas, even when low-complexity FDE receivers are employed.

Single-carrier with frequency-domain equalization (SC-FDE) modulations are a natural choice for the uplink of massive multiple intput, multiple output (MIMO) systems, due to the low peak-to-average power ratio (PAPR) of their signals. However, equalization techniques should avoid matrix inversions. On the other hand, energy-efficient analog-to-digital converters (ADCs) should be used in each receive branch to reduce the power consumption and the implementation complexity, which might lead to strong nonlinear distortion effects. In this work, we present the analytical performance evaluation of nonlinear massive MIMO systems based on SC-FDE systems employing low-complexity, iterative FDE receivers1.

In this paper, we consider massive multiple input multiple output systems that employ linear precoding techniques and are impaired by transmitter-side nonlinearities and channel estimation errors. We present an analytical method for the statistical characterization of the transmitted signals and develop a general framework to obtain analytically, the performance of different precoding techniques under the assumption of imperfect channel state information and for different types of nonlinearities. Regardless of the particular nonlinear transmission scenario, it is shown that the performance penalty associated to the nonlinear distortion can be tolerable if the number of transmit antennas is greater than the number of independent data streams.

It is well-known that Massive MIMO systems (Multiple-input multiple-output) have high potential for future wireless broadband systems. Massive MIMO (m-MIMO) relies on spatial multiplexing, and for that reason the base station needs a precise channel knowledge at uplink and downlink. Pilots can be used for channel state information (CSI) estimation, but common estimation processes imply a matrix inversion which can be a heavy computational process for m-MIMO system where the number of antennas used in the communication is very high. To alleviate computational requirements, reduce latency and to improve battery life capacity of mobile devices matrix inversion operations should be avoided. Having in mind these constrains, a new channel estimation method based on Zadoff-Chu (ZC) sequences is presented here, that achieves similar or better performance than least squares (LS) or minimum mean-Square Error (MMSE) channel estimators. It is also presented a set of performance results that sustain our assumption.

Massive multiple input, multiple output (MIMO) schemes have been considered to support the physical layer of 5G systems and its combination with single-carrier with frequency-domain equalization (SC-FDE) schemes is particularly interesting for the uplink. However, the receiver complexity increases with the number of antennas, and it is important to have low complexity massive MIMO schemes. In this paper we consider the receiver design for the uplink of massive MIMO schemes where SC-FDE techniques are employed by the user terminals. To achieve this, we empoly low resolution analog-to- digital converters (ADCs) at each receive branch of the BS, combined with low complexity FDE techniques. It is shown that, although the nonlinear distortion levels inherent to the use of low resolution ADCs can be very high, we can have excellent performance, even with low complexity FDE receivers, provided that the number of receiver antennas is higher than the number of user terminals.

Next generation of mobile communication systems must support astounding data traffic increases, higher data rates and lower latency, among other requirements. All these requirements should be met while assuring energy efficiency for mobile devices and base stations. Most likely, the future 5G systems will include massive MIMO (Multiple Input Multiple Output) schemes with dozen or hundreds of antennas enabling beamforming while operating in the millimeter wave spectrum. As soon as the millimetric wave propagation difficulties are overcome, the full potential of massive MIMO structures can be tapped. This paper presents a transmission system with bi-dimensional antenna arrays, based on a double layer structure combining beamforming with a multi-branch power amplification that achieves simultaneously power efficiency in amplification of multilevel constellations without penalties on system performance when compared with common schemes using only beamforming.

Orthogonal frequency division multiplexing (OFDM) schemes have high envelope fluctuations and peak- to-average power ratio (PAPR), making them very prone to nonlinear distortion effects, which can affect significantly the performance when conventional receivers are employed. However, it was recently shown that strong nonlinear distortion effects on OFDM signals do not necessarily lead to performance degradation. In fact, nonlinear OFDM schemes can outperform linear ones when optimum maximum likelihood (ML) receivers are employed. In this paper, we considered OFDM schemes with strong nonlinear distortion effects and we proposed a low- complexity detection scheme able to approach the optimum ML performance. Our technique is based on the fireworks algorithm (FWA) and allows excellent trade-offs between performance and complexity.

The use of massive multiple input multiple output (MIMO) techniques combined with orthogonal frequency division multiplexing (OFDM) schemes in the downlink of broadband wireless systems is considered in this paper. It is assumed that the transmitter supports up to R simultaneous links and employs T≫R antenna elements, and we study the feasibility of very low complexity transmitters with low resolution digital-to-analog converters (DACs). It is shown that the nonlinear distortion levels for detection purposes decrease with the ratio T/R, which means that we can employ very low resolution quantizers.

In this paper we consider massive multiple input multiple output (MIMO) schemes where each transmitter branch has very low resolution digital to analog converters (DACs). We present an analytical method for obtaining the statistical characterization of the transmitted signals that allows a simple and accurate performance evaluation even when there is imperfect channel state information (CSI). It is shown that, although the performance penalty associated to the use of low resolution DACs can be considerable, it can be mitigated by employing more antennas at the transmitter than at the receiver.

In this paper, we consider a T \times R massive MIMO system with up to R simultaneous links (T>> R>>1) and nonlinear amplifiers in each transmitter branch. The impact of a nonlinear amplification is analytically studied and accurate expressions for the spectral characterization of the transmitted signals, as well as the bit-error-rate (BER) are derived. The accuracy of our analysis is validated by simulations. It is also shown that the nonlinear distortion levels for detection purposes decrease with the ratio T/R, which means that the impact of the nonlinear distortion effects on the system's performance can be alleviated provided that T>>R. Therefore, the combination of massive MIMO systems with nonlinear, highly efficient power amplifiers can be an interesting option for low-cost, broadband wireless systems.

The use of massive multiple input multiple output (MIMO) techniques combined with orthogonal frequency division multiplexing (OFDM) modulations is being proposed for future broadband wireless systems such as 5G cellular networks. However, although these combination brings large capacity gains, it also suffers from two important drawbacks: the high implementation complexity inherent to the large number of antenna elements and the high sensitivity to nonlinear distortion effects due to the large peak-to-average power ratio (PAPR) of OFDM signals. In this paper, we study the capacity of massive MIMO-OFDM systems with strong nonlinear distortion effects. We derive theoretical expressions for the channel capacity considering different downlink scenarios where a base station with T nonlinear transmitting branches communicate with R receive antennas. It is shown that, although nonlinear distortion effects can reduce substantially the system capacity, this capacity loss can be reduced by increasing the number of transmit antennas (i.e., by using T≫ R).1.

Multiple input multiple output (MIMO) schemes combined with orthogonal frequency division multiplexing (OFDM) modulations are widely employed in broadband wireless systems. However, the transmitted signals are prone to nonlinear distortion effects due to their high envelope fluctuations. In this paper, we present a simple analytical method for studying nonlinear effects in MIMO-OFDM systems. Our method allows accurate results for the power spectral density (PSD) of the transmitted signals, as well as the signal- to- interference (SIR) levels at the detection level. It is shown that the robustness to nonlinear distortion effects increases with the number of transmit antennas (for a fixed number of independent data streams).

The combination of massive multiple input multiple output (MIMO) schemes with orthogonal frequency division multiplexing (OFDM) modulations are being considered for 5th generation (5G) wireless broadband systems. However, both the high complexity associated to massive MIMO transceivers as well as the amplification problems of OFDM constitute important drawbacks of such combination. In this paper we propose a reduced complexity, energy efficient massive MIMO OFDM scheme. For that purpose, we consider a maximal ratio transmit (MRT) technique combined with a simple time-domain clipping operation. We present an analytical method for obtaining both the power spectral density (PSD) of the transmitted signals, as well as the signal-to-interference (SIR) at the detection level. It is shown that having a large number of transmit antennas not only provides the known power and capacity gains, but also increases the robustness to nonlinear distortion effects.

In wireless communication systems privacy among users contents is a crucial security requirement such as power and spectral efficiency. Spectral and power efficiencies are attainable with power efficient transmission schemes where constellations are decomposed into several BPSK (Bi Phase Shift Keying), being each one amplified and transmitted independently by an antenna. Besides that, due to constellation shaping on the desired transmission direction performed by these transmitters, some kind of physical layer security is assured as well. Under this approach, security is achieved by constellation shaping of the transmitted constellation since each user must know the transmitter configuration parameters associated to the constellation shaping, i.e., the direction in which the constellation is optimized, otherwise the received data would be meaningless. Commonly, security is assured by encrypted algorithms implemented by higher layers, such as private and public encrypted keys. However, system security can be improved with physical layer security schemes since they can be complemented with other security schemes from higher layers. Simulation results show the effectiveness of the proposed approach.

The study of nonlinear effects on orthogonal frequency division multiplexing (OFDM) signals can be made taking advantage of their Gaussian characteristics. This is relatively simple and straightforward for smooth, polynomial nonlinear characteristics, since the required number of intermodulation products (IMPs) can easily be obtained. However, the situation is more complex for clipping characteristics, since they are not differentiable and their polynomial approximation is a function of the clipping level. In this paper we consider the effects of clipping on sampled OFDM signals and we define an equivalent nonlinear characteristic that is polynomial, with small degree, whose output has the same spectral characteristics of a clipping operation. Our results show that we can use this equivalent nonlinear characteristic for studying the impact of the effects of clipping on sampled OFDM signals, allowing very accurate estimates of the power spectral density (PSD) of the output signals, as well as the corresponding signal-to-interference ratio (SIR). It can also be employed to obtain the performance of optimum receivers for clipped OFDM.

Orthogonal frequency division multiplexing (OFDM) modulations are very popular in wireless communications due to their flexibility and good performance over severely time-dispersive channels. However, OFDM signals have large peak-to-average power ratio (PAPR), which leads to amplification problems. Constant envelope OFDM (CE-OFDM) techniques use an OFDM signal to modulate the phase of a given carrier, allowing a 0 dB PAPR, with all inherent advantages. However, the phase modulator is a nonlinear device that can lead to performance degradation. In this paper, we take advantage of the Gaussian characteristics of OFDM signals to characterize analytically CE-OFDM signals, both in terms of power spectral density (PSD) and optimum asymptotic performance.

Mobile communication systems must support multiple users achieving at same time privacy. Thus security is a crucial requirement to assure privacy of the data transmitted from and for each user. Several solutions can be adopted to implement security. Commonly they are based on encrypted algorithms implemented by higher layers, such as private and public encrypted keys. Also security can be implemented at physical layer and complemented with other security schemes from higher layers. Directivity introduced at the transmitted constellation can be also used to assure some kind of physical layer security. This is the case of power efficient transmission schemes where constellations are decomposed into several BPSK (Bi-Phase Shift Keying), being each one amplified and transmitted independently by an antenna. It can be shown that privacy among user contents is achieved under this approach, since each user must know the transmitter parameters associated to the constellation shaping, i.e., the direction in which the constellation is optimized, otherwise receives a meaningless data. If transmitter parameters are unknown, mutual information will be null which means a secrecy level of 100%. Simulation results show the effectiveness of the proposed approach.

Orthogonal frequency division multiplexing (OFDM) schemes are being considered for both optical fibre and wireless optical communications mainly because their ability to cope with the high inter‐symbol interference levels associated to dispersive channels. DC biased optical OFDM (DCO‐OFDM) is a promising technique that meet the particularities of intensity modulated OFDM schemes. To reduce the required DC component OFDM signals need to be submitted to an asymmetric clipping. However, the resulting non‐linear distortion effects can lead to significant performance degradation and/or undesirable out‐of‐band radiation. Therefore it is important to evaluate the impact of an asymmetric clipping on DC‐biased optical OFDM signals. In this work, the authors take advantage of the Gaussian‐like nature of OFDM signals and they study analytically the impact of an asymmetric clipping operation in the performance of DCO‐OFDM schemes. The authors present accurate theoretical expressions for the power spectral density and the non‐linear distortion at the subcarrier level, which can be used for obtaining the bit error rate of conventional receivers. The authors also study the optimum performance of these non‐linear OFDM schemes and the authors show that for an optimum maximum likelihood detection the non‐linear distortion instead of leading to performance degradation can actually improve it, even outperforming conventional, linear OFDM schemes.

It is widely recognized that orthogonal frequency division multiplexing (OFDM) signals are very prone to nonlinear distortion effects, which can lead to significant performance degradation. However, recent results have showed that nonlinear distortion effects do not necessarily mean performance degradation and can actually lead to performance improvements relative to conventional linear OFDM schemes. In this paper, we consider the effects of bandpass memoryless nonlinear devices on OFDM signals and study the optimum asymptotic performance for both nondispersive channels and severely time-dispersive channels. We present analytical methods for obtaining the Euclidean distance between two OFDM signals that are subjected to different nonlinear characteristics. These results are then employed for obtaining the average asymptotic gain relative to conventional linear OFDM schemes in nondispersive channels and the gain distribution for severely time-dispersive channels with Rayleigh-distributed fading on the different multipath components. Our analytical results, which are shown to be very accurate, indicate that the optimum detection of OFDM schemes with strong nonlinear distortion effects allows significant gains when compared with conventional linear OFDM schemes.

Amplify-and-forward relay techniques combined with OFDM (Orthogonal Frequency Division Multiplexing) modulations are being considered for broadband wireless systems since they allow simple relay systems suitable for severely time-dispersive channels. However, since they rely in OFDM-based signals, these techniques are prone to nonlinear distortion effects. In this paper we consider OFDM- based amplify-and- forward relay systems with strong nonlinear distortion effects at both the transmitter and the relay node and we present a simple, but accurate, analytical method for characterizing the distortion levels at the subcarrier level.

In this paper, we present a theoretical analysis for obtaining the minimum Euclidean distance between two nonlinearly distorted OFDM (Orthogonal Frequency Division Multiplexing) signals. This analysis takes advantage of the Gaussian behavior of OFDM signals with a large number of subcarriers and is a key step to get the asymptotic gain of the optimum receiver. We consider both polar and Cartesian memoryless nonlinearities, showing that in general the optimum performance of nonlinear OFDM is better than linear OFDM.

It was shown recently that we can decompose multilevel constellations as the sum of constant-envelope components which can be amplified and transmitted by separate antennas, allowing power-efficient transmitters, together with directivity at the constellation level without changing on the radiation pattern associated to the set of antennas. However, errors in the direction estimates can lead to substantial performance performance degradation since the constellations seen at the receiver can be substantially distorted. In this paper we present an improved receiver that is designed taking into account constellation distortion effects inherent to errors in direction estimates. It is shown that these "smart" receivers, optimized taking into account the apparent constellation at the receiver side can substantially outperform conventional receivers that assume that assume undistorted constellations.

Nonlinear phase distortion can severely degrade the performance of optical signals, both in terms of spectral characteristics and BER (Bit Error Rate) performance. This is especially serious for multicarrier optical signals. In this paper we study analytically the impact of nonlinear phase distortion on Coherent Optical Orthogonal Frequency Division Multiplexing (CO-OFDM) schemes with large number of subcarriers. The nonlinearly-distorted signal can be decomposed as the sum of useful and self-interference components. Conventional CO-OFDM implementations treat the self-interference as an undesirable noise term that leads to performance degradation. However, it is shown that this distortion component has information on the transmitted signals that can be employed to improve the performance. We show that the nonlinear phase distortion can lead to performance improvements relatively to the ideal linear case, contrary to what one could expect. We also present a sub-optimum receiver that is able to take advantage of this potential performance improvement.

Multilevel modulations can ensure high spectral efficiency but at cost of energy efficiency and envelope fluctuations that can compromise the amplification efficiency at the transmitter. Due to strict power and bandwidth constrains it is also desirable to adopt energy efficient amplification combined with powerful equalizer to reduce interference impact. In this paper we propose a method for amplification of multilevel constellations compatible with grossly nonlinear amplifiers based on the constellation's decomposition as a sum of BPSK (Bi-Phase Shift Keying) sub-constellations. Our approach relies on an analytical characterization of the mapping rule were the constellation symbols are written as a linear function of the transmitted bits. This analytical method is then employed to design iterative receivers such as IB-DFE (Iterative Block Decision Feedback Equalization), that can cope with higher sensitivity to ISI effects (InterSymbol Interference) of the resulting high order constellations. It is also defined a pragmatic method for designing IB-DFE receivers that can be employed with any constellation. It is also shown that the resulting method for designing an SC-DFE receiver (Single Carrier-Decision Feedback Equalization) does not require a significant increase in system complexity and can be used for the computation of the receiver parameters for any constellation.

In this paper we consider the optimum detection of OFDM (Orthogonal Frequency Division Multiplexing) signals with strong nonlinear distortion effects. It is shown that the optimum performance with strong nonlinear distortion effects is not as bad as one might expect and can even be better than the performance with conventional, linear transmitters. To achieve these excellent performances we should employ receivers able to take advantage of the information associated to transmitted data symbols that is inherent to the nonlinear distortion component, in opposition to traditional OFDM implementations where nonlinear distortion effects are regarded as an undesirable noise-like component. We study the achievable gains of the optimum receiver both analytically and by simulation. Since the complexity of optimum receivers is extremely high when we have nonlinear distortion effects, even for OFDM signals with a small number of subcarriers, we propose several sub-optimum receivers and evaluate their performance. Our sub-optimal receivers allow remarkable performance improvements, being able to reduce significantly the gap between the optimum performance and the performance of typical OFDM receivers.

Offset signals can be designed to have reduced envelope fluctuations allowing high amplification efficiency. When combined with SC-FDE (Single-Carrier with Frequency-Domain Equalization) schemes, they become excellent candidates for broadband wireless systems with severe power constraints. However, the performance of these modulations with conventional FDE receivers is sub-par, even for receivers specifically designed for offset modulations. This is especially serious when large offset QAM (Quadrature Amplitude Modulation) constellations are used. In this paper we consider SC-FDE schemes combined with offset modulations and study the reason behind the poor performance of conventional FDE implementations. We also present pragmatic FDE receivers for offset modulations that have low complexity and excellent performance.

This paper presents a Medium Access Control (MAC) protocol solution designed to properly handle collisions when in the presence of a multi-packet detection receiver for Single-Carrier (SC) modulations with Frequency-Domain Equalization (FDE). It is considered an iterative frequency-domain receiver that jointly performs equalization, multi-packet separation and channel decoding operations, for up to Qmaxmobile terminals transmitting in one slot. In this work, it is proposed and evaluated a p-persistent Network Diversity Multiple Access (NDMA) random MAC protocol designed to cope with a total number of mobile terminals J, for a maximum decoding capability of Qmaxsimultaneous packets. An accurate analytical model is presented to optimize two different scenarios: in the first one, a saturated network is considered and it is determined the packet transmission probability that maximizes the uplink throughput; the second represents a non-saturated network and the goal is to compute the optimal transmission probability associated to each mobile terminal that minimizes the packet transmission delay. In the end, analytical results obtained through physical and MAC layer simulations are discussed.

The high envelope fluctuations of OFDM signals (Orthogonal Frequency Division Multiplexing) make them very prone to nonlinear distortion effects. In typical OFDM implementations the nonlinear distortion component is regarded as a noise-like term that leads to performance degradation. To achieve optimum performance we should employ a ML (Maximum Likelihood) receiver where we take into account all the information associated to the transmitted signals that is in the nonlinearly-distorted signal. Although the performance of an ML is substantially better than traditional receivers, its complexity is prohibitively high. In this paper we consider nonlinearly distorted OFDM schemes and we present sub-optimum receivers that try to approach the ML performance. It is shown that, contrarily to what was expectable, the ML performance with nonlinear transmitters can be better than with ideal, linear transmitters. Our suboptimal receivers allow remarkable performance improvements, being able to reduce significantly the gap between the ML performance and the performance of typical OFDM receivers.

Offset modulations such as OQPSK (Off- set Quaternary Phase Shift Keying) and OQAM (Offset Quadrature Amplitude Modulation), combined with SC- FDE (Single-Carrier with Frequency-Domain Equalization) are an interesting choice for broadband wireless systems, that require low-complexity transmitters and efficient power amplification. However, the FDE perfor- mance can be rather poor due to interference between the in-phase and quadrature components at the sampling instants, especially when large constellations are employed. In this paper we propose pragmatic iterative and non- iterative FDE designs for offset modulations. Our performance results show that both iterative and non-iterative pragmatic FDE receivers have excellent performance, even for large OQAM constellations.

Orthogonal frequency division multiplexing (OFDM) schemes are the choice modulation for broadband wireless broadcast and multicast systems. However, OFDM schemes have important limitations such as high-envelope fluctuation of the transmitted signals and its sensibility to carrier frequency errors. When these limitations prove critical the authors consider single-carrier frequency-domain equalisation (SC-FDE) schemes, that allow much higher power efficiency because of lower envelope fluctuations. The overall performance can be further improved if the conventional linear FDE is replaced by an iterative FDE such as an iterative block decision feedback equaliser (IB-DFE). Conventional IB-DFE are usually designed for a quadrature phase shift keying constellation. However, it is strongly recommended that broadcast and multicast systems employ hierarchical constellations with several classes of bits with different error protection. In this study the authors consider the use of SC-FDE schemes combined with IB-DFE receivers in broadband wireless broadcast and multicast systems. The authors emphasise the advantages of these schemes and the authors present IB-DFE designs suitable for hierarchical constellations with several classes of bits with different error protection.

It is widely accepted that SC-FDE (Single-Carrier with Frequency-Domain Equalization) is an excellent candidate for broadband wireless systems, especially when an efficient power amplification is intended. If grossly nonlinear power amplifiers are employed, conventional QPSK (Quaternary Phase Shift Keying) or QAM (Quadrature Amplitude Modulation) modulations should be replaced by offset modulations such as OQPSK (Offset QPSK) and OQAM (Offset QAM). In fact, offset signals have much lower dynamic range than non-offset signals, with OQPSK signals being able to have an almost constant envelope. This paper considers frequency-domain receiver design for OQPSK and OQAM schemes. It is shown that FDE (Frequency Domain Equalization) designed for non-offset modulations are not suitable for offset modulations due to the residual IQI (In-phase/Quadrature Interference), i.e. the interference between in-phase (I) and quadrature (Q) components at the sampling instants. Therefore, we propose several FDE designs where there is no IQI in the sampling instants as well as iterative FDE receivers with IQI cancellation. Our receivers have excellent performance, with the linear designs significantly outperforming linear FDE for non-offset modulations and the iterative designs with performance close to the MFB (Matched Filter Bound).

In this paper we consider OFDM (Orthogonal Frequency Division Multiplexing) and SC-FDE (Single-Carrier with Frequency-Domain Equalization) schemes where the channel is estimated with the help of training blocks multiplexed with data blocks. It is shown that SC-FDE schemes need more accurate channel estimates than OFDM schemes with the same constellation. To improve the performance, we consider iterative receivers with joint detection and channel estimation. Our receivers allow significant performance improvements, for both OFDM and SC-FDE, with OFDM slightly outperforming SC-FDE.

In SFN (Single Frequency Network) broadcasting systems the equivalent CIR (Channel Impulse Response) can be very long but typically it has a sparse nature. This means that the equivalent CIR has several clusters of paths, each one associated to the CIR between a given transmitter and the receiver. These clusters have several multipath components and are typically well separated in time. In this paper we consider OFDM-based (Orthogonal Frequency Division Multiplexing) broadcasting systems with SFN operation and we propose an efficient channel estimation method that takes advantage of the sparse nature of the equivalent CIR. For this purpose, we employ low-power training sequences and we employ an iterative receiver with joint detection and channel estimation. The receiver can know the location of the different clusters that constitute the overall CIR or not. Our performance results show that we can have excellent performance, close to the performance with perfect channel estimation, even with employ low-power training blocks and also for the case where the receiver does not know the location of the different clusters that constitute the overall CIR.

SC-FDE (Single Carrier Frequency-Domain Equalization) modulations are known to be suitable for broadband wireless communications due to their robustness against severe time-dispersion effects and the relatively low envelope fluctuations of the transmitted signals. In this paper we consider the use of SC-FDE schemes in broadcast and multicast systems with SFN (Single Frequency Network) operation where we do not have perfect carrier synchronization between different transmitters. We study the impact of different CFO (Carrier Frequency Offset) between the local oscillator at the receiver and the local oscillator at each transmitter. We also propose receiver structures able to reduce the performance degradation caused by different CFO at different transmitters. Our receivers can be regarded as modified turbo equalizer implemented in the frequency-domain, where a frequency offset compensation is performed before the iterative receiver.

This paper considers the use of several grossly nonlinear amplifiers for an efficient amplification of Mary-OQAM signals (Offset Quadrature Amplitude Modulation) with FDE (Frequency Domain Equalization) and the impact of phase and gain unbalance between different amplifiers. It is shown that for some 16-QOQAM signals it is possible to employ two grossly nonlinear amplifiers to perform an amplification that is formally equivalent to a linear amplification. However, the phase and gain of these amplifiers need to be carefully selected, since we can only accept gain errors below 10% and phase errors below 0.1 rad (roughly 6 degrees).

In this paper we consider SC-FDE schemes (Single-Carrier with Frequency-Domain Equalization) employing multi-resolution hierarchical constellations and we consider the design of turbo receivers, implemented in the frequency-domain, for systems where we have different error protections. We present a versatile method to design the receiver that is suitable to any constellation.

This paper considers the design of linear frequency-domain receivers with oversampling for OQPSK (Offset Quaternary Phase Shift Keying) signals. Based on a receiver structure able to eliminate the residual interference between in-phase (I) and quadrature (Q) components at the sampling instants, a method that minimizes the overall residual ISI (Inter-Symbol Interference) plus IQI (I-Q interference) is proposed. Therefore, this receiver significantly outperforms previous frequency-domain OQPSK receivers.

Nonlinear operations are usually regarded as something undesirable that should be avoided in digital communication systems since they lead to performance degradation, namely spectral widening and decrease in power efficiency, for a given modulations. In this paper we show that the adoption of appropriate nonlinear devices can allow an improvement on the performance of some modulation schemes, either for OQPSK-type schemes (Offset Quadrature Phase Shift Keying) or TC-OQAM (Trellis Coded Offset Quadrature Amplitude Modulation). We present an analytical characterization of these signals after a bandpass memoryless nonlinear device. It is shown that an encoder followed by a nonlinear amplifier can be viewed as a serial concatenated coding scheme and the coding gain of such scheme can be improved relatively to the linear transmission scheme.

SC modulation (Single-Carrier) with FDE (Frequency-Domain Equalization) combined with iterative (turbo) FDE schemes has excellent performance in severely time-dispersive channels, making it a promising candidate for future broadband wireless systems. In fact, it was observed that the performance can be close to the MFB (Matched Filter Bound). In this paper we consider a class of iterative FDE schemes and we study the impact of the number of multipath components and the diversity order on its performance. It is shown that for a high number of separable multipath components the asymptotic performance approaches the MFB, even without diversity. When we have diversity the performance approaches the MFB faster, even when we have just a small number of separable multipath components.

This paper considers frequency-domain receiver design for OQPSK-type schemes. Since the I-Q interference can lead to significant performance degradation we propose an FDE design (frequency-domain equalization) where there is no I-Q interference in the sampling instants for the predominant OQPSK constellation. We also propose iterative FDEs with I-Q interference cancelation, as well as the cancelation of unintended OQPSK components.

The traditional approach to cope with collisions is to discard the packets involved in it and to ask for their retransmission. However, since the signal associated to a collision has important information concerning the packets involved, we can efficiently resolve collisions with proper retransmissions. In this paper we consider severely time-dispersive channels and we propose a technique that allows efficient packet separation. We employ SC-FDE schemes (single-carrier with frequency-domain equalization) and we propose an iterative frequency-domain multi-packet detection. Since our technique requires uncorrelated channels for different retransmissions, we also propose an SP technique (Shifted Packets) for retransmissions in fixed channels. Since the total number of transmissions is equal to the number of packets involved in the collision (even when the channel remains fixed for the retransmissions), our technique allows high throughputs. The analysis of our detection technique combined with the NMDA (network-assisted diversity multiple access) MAC (medium access control) protocol shows significant throughput and delay improvements for low Eb/N0values compared to a TDMA (time division multiple access) approach, where collisions are avoided. This technique is particularly appealing for the uplink of broadband wireless systems, since we consider SC-FDE schemes and the complexity is concentrated in the receiver. By employing the SP scheme we can use the same channel for the retransmissions, with only a small performance degradation.