Jeebak Mitra

Convolutionally Coded Transmission over Markov-Gaussian Channels: Analysis and Decoding Metrics

It has been widely acknowledged that the aggregate interference at the receiver for various practical communication channels can often deviate markedly from the classical additive white Gaussian noise (AWGN) assumption due to various ambient phenomena. Moreover, the physical nature of the underlying interference generating process in such cases can lead to a bursty behaviour of the interfering signal, implying that it is highly likely that consecutive symbols are affected by similar noise levels. In this paper, we devise and analyze detection techniques, in conjunction with a convolution code, for such interference channels that possess non-negligible memory by considering optimum and sub-optimum decoding metrics. In particular the inherent memory in the noise process is modeled as a first-order Markov chain, whose state selects the variance of the instantaneous Gaussian noise, leading to a Markov-Gaussian channel model. Analytical expressions are obtained for the cut-off rate, which is an ensemble code parameter, and the bit error rate for a convolutionally coded system, that are subsequently employed for an extensive evaluation of the various metrics considered. Furthermore, the interleaving depth is considered as a design parameter and its effect on performance is analyzed over a range of noise scenarios.

Robust Detectors for TH IR-UWB Systems with Multiuser Interference

It has been recently established that multiuser interference (MUI) in time-hopping impulse-radio ultra-wideband (TH IR-UWB) systems is highly impulsive and hence the conventional matched filter (MF) receiver based on the assumption of Gaussian MUI is no longer optimum. In this paper, two novel receiver structures for detecting TH IR-UWB signals in the presence of MUI are proposed. The first detector is based on a two-term approximation for the distribution of the interference process and noise and the second uses a non-linear penalty function for the decision statistic. Simulation results show that the proposed detectors perform considerably better than the conventional MF receiver and also outperform a recently devised soft-limiting detector in MUI limited environments.

Design and analysis of robust detectors for TH IR-UWB systems with multiuser interference

In this letter, we design and analyze the performance of single-user-type non-linear detectors that are able to cope with the impulsive nature of multiuser interference (MUI) in timehopping impulse-radio ultra-wideband (TH IR-UWB) systems. We collectively refer to these detectors as robust detectors. We first propose two novel detectors and then derive semi-analytical expressions for the bit-error rate (BER) of TH IR-UWB with general robust detection. The evaluation of these expressions greatly facilitates the optimization of detector parameters and provides insight into the effects of MUI. A performance comparison shows that (1) robust detection significantly improves performance over conventional detection in the presence of MUI, (2) the parameters for various parametric robust detectors can be chosen to be constant over many transmission scenarios with only little performance degradation compared to using the optimal parameter value, and (3) the proposed two-term detector, which requires a modest amount of parameter estimation, achieves consistently the best performance.

GEN01-3: Robust Decoding for Channels with Impulse Noise

Data transmission over power lines is an attractive alternative to well-established wireline and wireless communication technologies. One of the main challenges in accomplishing reliable power-line communication (PLC) is channel impairment through electromagnetic interferences, or so-called impulse noise. In this paper, we consider transmission over impulse-noise channels for a typical narrowband system architecture employing convolutional codes and Viterbi decoding. We present different decoding metrics, including new designs adopted from the multiuser detection literature, and we derive expressions for cutoff rate and bit-error rate (BER) performances of the corresponding decoders. These expressions are amenable for quick numerical evaluation and thus, constitute a valuable tool for decoder optimization and performance comparison. Our numerical and BER simulation results show that one of the proposed metrics enables robust decoding without knowledge of the statistic of the impulse noise with a performance close to that of optimum decoding, which relies on the noise statistic. It is further highlighted that, different from transmission over the Gaussian-noise channel, quadrature detection is beneficial in case of real-valued modulation and passband transmission over impulse-noise channels.

Coded narrowband transmission over noisy powerline channels

Powerline communication (PLC) channels present one of the harshest environments for efficient data transmission. In particular, asynchronous impulsive noise has been known to be one of the strongest impediments to error-free transmission over PLC channels. Most work aimed at ameliorating the degradation due to impulsive noise considers the noise impulses to be independent. In a realistic PLC channel, however, this is rarely true. This paper is concerned with the performance limits on communication over powerline channels where the noise has memory and is modeled using a partitioned Markov chain (PMC) that has been found to be well suited to describe the bursty nature of impulses of the low voltage PLC channel. In particular, expressions are derived for the cutoff rate and the bit error rate of a convolutionally coded narrowband system. They are then verified by comparing with simulation results employing typical PLC parameters, proving the utility of the expressions as a design tool.

On joint estimation and decoding for channels with noise memory

In this letter we consider coded transmission over interference channels where the interference occurs in bursts and hence is considered to be impulsive. The bursty nature of the interference leads to memory in the overall noise process which is modeled as a 2-state Markov chain. Recent work on coded transmission for such channels has proposed decoding techniques assuming that perfect knowledge of the interference statistics are available at the receiver. In this work, we aim at completing the picture by proposing a novel algorithm that decodes without the knowledge of the interference statistics and highlighting the differences between the two cases.

Comparison of detectors for multiple-access interference mitigation in TH-IR UWB

Multiple access interference (MAI) in time-hopping impulse-radio ultra-wideband (TH IR-UWB) systems has been shown to have a non-Gaussian distribution. Hence the conventional linear matched filter receiver is no longer the optimal solution, and a number of non-linear detectors that are ldquorobustrdquo to non-Gaussian impairments have been proposed in the recent past. Most of these detectors depend on one or more parameters and are thus parametric detectors. In this paper, we compare the performance of those detectors and investigate the dependency of the performance on proper parameter adjustment.

Sensing and suppression of impulsive interference

Cognitive radios are slated to be the next generation of smart transceivers that can dynamically sense and respond to its immediate radio frequency (RF) environment. It is highly likely that the RF environment will vary with time as various interferers come and go out of the range of the target receiver. This leads to the interference at the receiver being impulsive in nature, which if not properly handled can cause irrecoverable damage to the transmitted data. The traditional cognitive radio would, in such a scenario, decide against transmitting when a harmful interferer is present in the vicinity. In this work, we investigate methods to mitigate the effects of such interference through intelligent signal processing at the receiver such that throughput can be greatly enhanced. We introduce receiver structures for the more practical scenario of temporally correlated interference and quantify the achievable gains when simple yet effective interference suppression methods are applied at the receiver.

Serial concatenation of simple linear block codes and differential modulations

In this paper, binary linear block codes with very low encoding and decoding complexity (therefore, simple codes) are introduced as outer codes in a serially concatenated coding scheme employing multilevel differential phase-shift keying, i.e., differential modulations, as inner codes. Such a concatenated coding scheme is particularly suitable for power- and bandwidth-efficient transmission over channels with phase ambiguities or completely unknown phase at the receiver. Using extrinsic information transfer chart based analysis and optimization, the performance of the new concatenated codes is found to be within less than 1 dB of the pertinent capacity limit for the additive white Gaussian noise channel. This compares favorably with more complex benchmark schemes using e.g. outer low-density parity-check codes proposed recently in the literature. Simulation results confirm the predicted performance advantages for the proposed simple codes.

Evaluation of relative performance of product designs: A fuzzy dea approach to quality function deployment

The Fuzzy DEA approach is introduced in Quality Function Deployment (QFD) to evaluate relative performance of alternative product designs, when performance characteristics obtained from voice of customers are fuzzy, product design data are often limited, inaccurate and vague, and imprecise relationships exist among engineering and performance characteristics.