This special issue is devoted to Satellite Communication advances so as to gather the most important contributions in this field from academia and industry. In particular, the importance of this issue is also due to the fact that the last issue appearing on IEEE JSAC about SatCom is dated back to 2004 and therefore many important technology advances have been testified in the meantime, actually revolutionizing the satellite industry. In particular, important innovations have been brought to the design of both ground and space segments, spanning physical layers up to upper layers of the protocol stack.

This special issue is devoted to the satellite communications (SatCom) advances so as to gather the most important contributions in this field from academia and industry. In particular, the importance of this issue is also due to the fact that the last issue appearing on IEEE JSAC about SatCom is dated back to 2004 and therefore many important technology advances have been testified in the meantime, actually revolutionizing the satellite industry. In particular, important innovations have been brought to the design of both ground and space segments, spanning physical layers up to upper layers of the protocol stack.

This paper investigates linear precoding for bit-interleaved coded modulation (BICM) on additive white Gaussian noise (AWGN) channels. Concatenating a linear precoder as an inner encoder with an outer (convolutional) encoder produces a powerful code with a limited decoding complexity. Linear precoders are examined and optimized for two scenarios: using (i) a noniterative decoding strategy and (ii) an iterative decoding strategy under a perfect feedback assumption. The precoder design is based on an information-theoretical approach, on the one hand, and a pair-wise error probability (PEP) analysis, on the other hand. Both approaches render convenient precoder design rules. For a binary phase-shift keying (BPSK) signal set, the optimal precoders that result from these rules are also derived. Numerical results confirm the analytical findings and simulations illustrate the effectiveness of the approach.

This contribution deals with code-aided hypothesis testing for wireless digital receivers. We provide a theoretical justification for a hypothesis testing algorithm that was previously introduced in (Wymeersch et al., 2006) based on ad-hoc arguments. Contrary to conventional hypothesis testing methods, the algorithm from Wymeersch et al. exploits the code structure within the received signal and does not require any pilot symbols. By doing so, it allows to improve the bandwidth-efficiency of the transmission. The present contribution shows that, under mild conditions, the performance of the algorithm from Wymeersch et al. coincides with the performance of the optimal Maximum A Posteriori (MAP) hypothesis test. Computer simulations support this result.

This paper compares parametric and nonparametric channel estimation methods in a multipath environment. For both methods, the MSE on the received symbol pulse is shown to consist of a modeling error term and an estimation error term. We lower bound the latter term by computing the associated Cramer-Rao bounds. We point out that the MSE can be minimized by a proper selection of the number of parameters to be estimated. In addition, we present simulation results from practical estimators that confirm our theoretical analysis. Our numerical results indicate that the parametric method outperforms the nonparametric approach.

Spectrum-overlay scenarios for wideband multi-carrier (MC) systems bring new technical challenges that must be considered during the system design. In such a scenario, the receiver has to perform actions, such as channel estimation and synchronization, in the presence of possibly strong in-band interference. In this paper, we modify the iterative maximum likelihood joint frame synchronization and channel estimation algorithm for downlink MC-CDMA presented in (M. Guenach et al., 2006) to have better performance in the presence of narrowband interference. The performance of the modified algorithm is verified through computer simulations and compared with the original one.

In this contribution, we present a novel multiple-input multiple-output (MIMO) detection scheme for quadrature amplitude modulation (QAM). The proposed method is based on the slowest descent (SD) method and is suitable for both iterative and non-iterative detectors. Optimal MIMO detection entails maximizing or marginalizing a likelihood function over a very large set of vectors. Similar to other low-complexity detection schemes such as sphere decoding, the SD approach restricts the maximization/marginalization to a small set of candidate vectors. As the size of this set scales linearly with the number of transmit antennas, the SD method bears an exceptionally low complexity. Furthermore, simulation results indicate that the method achieves a close-to-optimal performance, provided that the diversity order is sufficiently high.

The use of linear precoding in bit-interleaved coded modulation with iterative decoding leads to a significant performance improvement on AWGN channels. In this contribution, we approach linear precoding from an information theoretical perspective. We compute the bitwise capacity as seen by each of the precoder inputs in order to investigate the impact of precoding for two scenarios: using (i) a non-iterative decoder and (ii) an iterative decoder under a perfect feedback assumption. When restricting to BPSK signaling, the bitwise capacity reveals simple precoding design rules for both scenarios. Simulation results confirm our analytical findings

This paper deals with joint data detection, synchronization and channel parameter estimation for direct-sequence code-division multiple-access systems over frequency-selective channels. In low-signal-to-noise ratio environments, conventional data-aided (DA) estimation algorithms may require an unacceptably large number of pilot symbols in order to obtain sufficiently accurate estimates of the channel and synchronization parameters. Especially, frequency offset estimation results in a significant loss in spectral efficiency. In this contribution, we consider several code-aided estimation schemes which can be incorporated in an iterative turbo detection scheme. We consider the expectation maximization algorithm, as well as the space alternating generalized expectation maximization algorithm as tools to develop code-aided estimation algorithms for a variety of scenarios. We pay special attention to the issue of computational complexity, and propose some complexity-reducing approximations. We show through computer simulations that the proposed code-aided estimation techniques considerably outperform their conventional DA counterparts.

This contribution deals with carrier frequency offset estimation for flat-fading multiple-input multiple-output (MIMO) channels. Both data-aided (using training symbols) and iterative code-aided (using the unknown coded symbols) estimation is considered. In both scenarios, maximum-likelihood (ML) frequency offset estimation involves solving a maximization problem with no closed-form solution. Since numerical calculation of the ML estimates is computationally hard, we derive a simple closed-form approximation. Simulation results indicate that the ML algorithm and the proposed reduced-complexity algorithm operate closely to the Cramer-Rao bound (CRB)

We describe a class of low-complexity channel parameter and frequency offset estimation algorithms, applied to a space-time bit-interleaved coded modulation (ST-BICM) scheme for burst-mode asynchronous DS-CDMA uplink transmission over frequency selective channels. Conventional estimation algorithms rely mainly on the presence of training symbols within the data stream, reducing both the energy- and bandwidth efficiency. Our main goal is to remove the need for long training sequences by exploiting the presence of an underlying error-correcting code. We derive an iterative estimation technique, based on the space alternating generalized expectation maximization (SAGE) algorithm, which iterates between detection and estimation stages. Performance results are illustrated through computer simulations

This paper deals with space-time detection and channel parameter estimation in a bit-interleaved coded modulation (BICM) scheme for asynchronous direct-sequence code division multiple access (DS-CDMA) transmission over frequency selective channels. Using a pilot sequence to obtain sufficiently accurate channel estimates for this system requires an unacceptably large number of pilot symbols. In this contribution, we consider several code-aided estimation schemes which can be incorporated in an iterative space-time (ST) turbo detection scheme. We point out that the EM algorithm, which has recently been addressed as a convenient tool for code-aided estimation, has some serious drawbacks because of the large number of parameters to be estimated. These can be resolved using the space alternating generalized expectation maximization (SAGE) algorithm. We show through computer simulations that the proposed low complexity ST receiver with SAGE estimation considerably outperforms conventional estimation schemes.

Code-aided parameter estimation is considered for bit-interleaved coded modulation with iterative decoding (BICM-ID). The computational overhead caused by the estimation is minimized by embedding the parameter estimation into the iterative decoding. It is apparent that the performance of the estimation depends on the availability of accurate symbol a posteriori probabilities (SAPPs). Since the decoder computes bit a posteriori probabilities (BAPPs), the SAPPs are traditionally computed through a straightforward bit-to-symbol probability conversion. For advanced mappings which are optimized for BICM-ID, however, the latter method does not result in reliable SAPPs. Therefore, a new method is proposed and simulation results verify that the resulting iterative estimator-detector yields a close to perfectly synchronized performance.

The paper deals with carrier frequency offset estimation for a flat-fading multiple-input multiple-output (MIMO) channel using a training sequence. The resulting maximum likelihood (ML) estimation entails solving a maximization problem with no closed-form solution. Since numerical calculation of the estimate is computationally hard, we propose a sub-optimal closed-form solution. In contrast with single-input-single-output (SISO) systems, however, self-noise arises in MIMO closed-form frequency offset estimation. Through proper training sequence design, we show how to avoid this self-noise and achieve a performance close to ML-performance and the Cramer-Rao bound (CRB).

This contribution proposes a new spatial mapping technique for bit-interleaved coded modulation (BICM) over multiple-input multiple-output (MIMO) channels. It is an extension of conventional BICM whereby the serial-to-parallel converter is replaced with a more general mapping strategy. Genie performance analysis of spatial mapping schemes results in very simple design criteria for the mapping function. Using an iterative detector, a significant performance gain is observed compared to a conventional BICM scheme, at no significant increase in computational complexity. The predicted performance gains are verified through computer simulations.