Bursty Impulsive Noise Research Articles

An LLR-Based Receiver for Mitigating Bursty Impulsive Noise With Unknown Distributions

An LLR-Based Receiver for Mitigating Bursty Impulsive Noise With Unknown Distributions

BNET: A Neural Network Approach for LLR-Based Detection in the Presence of Bursty Impulsive Noise

Ensuring reliable wireless communications is often a major challenge in environments characterized by bursty impulsive noise. Unfortunately, effective receivers adapted to this problem employ recursive techniques that are computationally complex and challenging to implement where computing and battery resources are limited. As a solution, we propose a computationally efficient neural network-based framework to approximate the log-likelihood ratio function produced by the Bahl Cocke Jelinek Raviv (BCJR) algorithm for single-carrier communication systems. The results show that our method can match the BCJR algorithm’s BER performance in many scenarios and even outperform it when some noise model parameters change over time.

Bursty Impulsive Noise Mitigation in NOMA: A MAP Receiver-Based Approach

Non-orthogonal multiple access (NOMA) was recently proposed as a promising multiple access technique that can enable improved spectral efficiency, massive connectivity, and low latency. In this letter, we analyze the performance of power domain NOMA in the presence of impulsive noise, which is encountered in various practical applications, such as smart-grid communications. Such noise is known to degrade overall system performance. Moreover, since the noise in question is bursty, it is an arduous task to detect the signals with conventional receivers. Therefore, this letter sheds light on the performance degradation and mitigation of bursty impulsive noise, modeled by a two-state Markov-Gaussian process, in the context of uplink NOMA systems. More specifically, we propose a maximum a posteriori (MAP) receiver combined with successive interference cancellation to mitigate the effect of impulsive noise on the detection of the different users' signals. Simulation results prove the effectiveness of the proposed MAP-based bursty impulsive noise mitigation scheme in terms of bit error rate.

Deep Learning for MMSE Estimation of a Gaussian Source in the Presence of Bursty Impulsive Noise

We develop a deep learning (DL) based Gaussian source estimation technique when the source is impaired by bursty impulsive noise. This noise is correlated in time, and hence estimating the source with a low-complexity algorithm is a challenging task. To address this challenge, we train a long short term memory (LSTM) based deep neural network (DNN) model offline with different bursty noisy observations and deploy the trained model in real-time. The trained model detects the noise state online and thus applies a linear minimum mean square error (LMMSE) method to estimate the source signal. To demonstrate the effectiveness of the proposed scheme, we compare its performance with baseline schemes. Simulation results reveal the effectiveness of the proposed estimation technique in terms of mean square error and computational complexity.

Mitigation Techniques for Impulsive Noise With Memory Modeled by a Two State Markov-Gaussian Process

Impulsive noise, a common impediment preventing the system from achieving error-free transmission, is significant in many wireless and power line communication environments. Although the performance of several mitigation techniques for orthogonal frequency division multiplexing (OFDM)-based multicarrier communication systems impaired by memoryless impulsive noise are widely acknowledged, we note that OFDM is outperformed by its single-carrier counterpart when the impulses are very strong and/or they occur frequently, which is likely to exist in contemporary communication systems including smart grid communications. On the other hand, many communication technologies used in the smart grid do not employ OFDM and likewise, the assumption of memoryless noise is not valid for such communication scenarios. Memoryless noise models cannot take into account one of the main features of the actual noise, i.e., the time-correlation among the noise samples. The aim of this article is to compare and analyze several mitigation techniques such as clipping, blanking, and combined clipping-blanking to mitigate the noxious effects of bursty impulsive noise for low-density parity-check coded single-carrier communication systems. Moreover, we propose a log-likelihood ratio (LLR)-based impulsive noise mitigation for the considered scenario. In this context, provided simulation results highlight the superiority of the LLR-based mitigation scheme over the clipping/blanking schemes.

A Novel Relay Selection Strategy of Cooperative Network Impaired by Bursty Impulsive Noise

Best relay selection (BRS) is crucial in enhancing the performance of cooperative networks. In contrast to most previous works, where the guidelines for BRS are limited to Gaussian noise, in this paper, we propose a novel relay selection protocol for a decode-and-forward cooperative network taking into account the bursty impulsive noise (IN). The proposed protocol chooses the $N$ th best relay considering both the channel gains and the states of the IN of the source-relay and relay-destination links. For this scheme, to obtain the state of IN, we propose a state detection algorithm using maximum a posteriori (MAP) detection. To analyze the performance of the proposed protocol, we first derive closed-form expressions for the probability density function (PDF) of the received signal-to-noise ratio assuming all the relays know the state of IN perfectly (genie-condition). Then, these PDFs are used to derive closed-form expressions for the bit error rate (BER) and the outage probability. Finally, we also derive the asymptotic BER and outage expressions to quantify the diversity benefits. We show that the proposed MAP-based $N$ th BRS protocol attains the derived genie-aided analytical results and outperforms the conventional relay selection protocol, optimized for the Gaussian case, and which does not take into account the IN memory.

BITFCM-OFDM scheme for power-line communication systems

Power-line communications (PLC) is viewed as a good potential solution for smart grid communications because of the established infrastructure around the world. However, power-line was not initially aimed to be used for continuous data communications. Hence, challenging problems may occur when the power-line channel is used as a communication channel. One of the major issues is the (bursty) impulsive noise in the power-line channel. This paper analyzes the system performance of the bit-interleaved time-frequency coded modulation (BITFCM) scheme with an optimal low complexity maximum likelihood (ML) decoding scheme in the orthogonal frequency-division multiplexing (OFDM)-based broadband PLC (BB-PLC) modems. The numerical results illustrate that this proposed scheme improves the system performance of the PLC modems under the bursty impulsive noise environment. Furthermore, the performance of the BITFCM-OFDM scheme can achieve up to 3 dB improvement at bit error rate (BER) of 10-4 in comparison with the bit-interleaved coded modulation (BICM)-OFDM and shows significant improvement compared with the coded OFDM (COFDM). Finally, a blanking nonlinear preprocessor is used at the receiver and further improvement of the BER performance is evaluated.

Comment on “Bayesian MMSE Estimation of a Gaussian Source in the Presence of Bursty Impulsive Noise”

In the above-titled letter by Alam et al. (see ibid., vol.22, no.9, p.1846–49, September 2018), the authors claimed to derive the minimum-mean-squared-error (MMSE) estimator for a scalar Gaussian source in the presence of correlated noise, modeled by a two-state Gauss-Markov process. In the proposed approach, first, the state of the noise process at time epoch k is determined according to the maximum a posteriori (MAP) decision rule based on the whole sequence of observations. Then, given the hard decision on the noise state, the corresponding MMSE estimator (i.e., conditional mean) is employed to estimate the source signal at time epoch k . However, performing estimation based on the hard decision of the MAP detector is not optimal in terms of minimizing the mean-squared-error (MSE). In this comment, we show that the optimal MMSE estimator for the considered problem is obtained as a weighted average of the MMSE estimators corresponding to distinct states of the noise probable at time epoch k and the weights are equal to the a posteriori probabilities of the noise state computed based on the whole sequence of observations.

Bayesian MMSE Estimation of a Gaussian Source in the Presence of Bursty Impulsive Noise

In this letter, we derive the minimum mean square error (MMSE) optimal Bayesian estimation (OBE) for a Gaussian source, in the presence of bursty impulsive noise, as essentially encountered within power substations. Clearly, it is observed that the presence of bursty impulsive noise makes the input–output characteristics of MMSE OBE non-linear. To handle the non-linearity, we propose a novel MMSE estimator, based on the detection of the unobservable states of the noise process, using the maximum a posteriori (MAP) detector. Resultantly, the proposed MAP-based MMSE estimator is shown to achieve the lower bound derived for the proposed scenario and outperform the various MMSE estimators that neglect the noise memory.

Analysis of Bursty Impulsive Noise in Low-Voltage Indoor Power Line Communication Channels: Local Scaling Behaviour

Power line communication (PLC) channels are prone to multipath propagation due to impedance mismatch, and impulsive noise whose characteristics are still not well established in the literature. Moreover, measurements show that this impulsive noise appear in bursts, non-Gaussian and cyclostationary and as such cannot be modeled as the convenient additive white Gaussian noise (AWGN). Transceivers optimized for AWGN may not necessarily perform well for the PLC noise. Therefore, investigating the characteristics of PLC noise is very important for accurate modeling of the same. This study presents multifractal analysis of bursty impulsive noise measured from power line networks from three different environments. We employ multifractal detrended fluctuation analysis, which is a well-developed multifractal analysis technique for non-stationary time series data and easy to implement to analyze measured noise data. Results show that power line noise exhibits both long-range dependence (LRD) and multifractal scaling behavior with different strengths depending on the environments where they were captured. The multiscaling behavior is due to long-range correlation inherent in the power line noise. The source of this local multiscaling behavior is determined by analysis a shuffled series of the original data captured from the power network. Multifractal analysis is able to show clearly both the strengths and frequency of occurrence of bursts occurring in PLC noise which can then be applied in accurate modeling of the noise. The significance of these results is that new power line noise models should be developed that captures both LRD and multifractal scaling for more accurate performance evaluation of power line communication systems. The existing noise models though able to replicate temporal dependence of PLC noise, are not able to capture this local scaling behavior which results show is inherent in PLC noise.

Optimal and Near-Optimal Detection in Bursty Impulsive Noise

In many practical scenarios, the ambient noise process is known to be impulsive. To combat this, several robust measures have been proposed in the literature. Most of them assume white noise processes, i.e., the noise samples are independent and identically distributed heavy-tailed random variables. However, noise is seldom white in practice and therefore exhibits memory. For impulsive noise, dependency among samples results in outliers clustering together. The process is thus impulsive and bursty. In our work, we employ stationary $\alpha$ -sub-Gaussian noise with memory order $m$ ( $\alpha$ SGN( $m$ )) to model bursty impulsive noise. The model is based on the multivariate $\alpha$ -sub-Gaussian ( $\alpha$ SG) distribution family and statistically characterizes adjacent samples from elliptical distributions. The latter assumption holds well for snapping shrimp noise found in warm shallow underwater channels. We investigate the performance of conventional robust detectors in $\alpha$ SGN( $m$ ) and also propose novel near-optimal detectors. The Neyman–Pearson (NP) approach for binary hypothesis testing is considered and extensive simulation results for the aforementioned detectors are offered. For all instances, we employ an $\alpha$ SGN( $m$ ) process whose parameters are tuned to snapping shrimp noise data sets. By incorporating good signal design rules, it is shown that there is a large performance gap between the new and conventional detectors in various impulsive regimes. Moreover, it is possible to derive a near-optimal detector if one only has information of the temporal statistics of the noise process.

Discrete Multi-Tone Digital Subscriber Loop Performance in the Face of Impulsive Noise

As an important solution to “the last mile” access, digital subscriber loops (DSLs) are still maintained in a huge plant to support low-cost but high-quality broadband network access through telephone lines. The discrete multi-tone (DMT) transmissions constitute a baseband version of the ubiquitous orthogonal frequency division multiplexing. While the DMT is ideally suited to deal with the frequency selective channel in DSL, the presence of bursty impulsive noise tends to severely degrade the transmission performance. In this paper, we analyze the statistics of impulsive noise and its effects on the received signals, with the aid of a hidden semi-Markov process. The closed-form bit error rate expression is derived for the DMT system for $Q$ -ary quadrature amplitude modulation under practical noise conditions and for measured dispersive DSL channels. Instead of relying on the simplified stationary and impulsive noise process, our noise model considers both the temporal and spectral characteristics based on the measurement results. The simulation results confirm the accuracy of the formulas derived and quantify the impact both of the impulsive noise and of the dispersive channel in DSL.

Performance Analysis of DF Cooperative Relaying Over Bursty Impulsive Noise Channel

In this paper, we consider the performance analysis of a decode-and-forward (DF) cooperative relaying (CR) scheme over channels impaired by bursty impulsive noise. Although the Middleton class-A model and the Bernoulli–Gaussian model give good results to generate a sample distribution of impulsive noise, they fail in replicating the bursty behavior of impulsive noise, as encountered for instance within power substations. To deal with that, we adopt a two-state Markov–Gaussian process for the noise distribution. For this channel, we evaluate the bit error rate performance of direct transmission (DT) and a DF relaying scheme using M-ary phase shift keying (M-PSK) modulation in the presence of Rayleigh fading with a maximum a posteriori (MAP) receiver. From the obtained results, it is seen that the DF CR scheme in bursty impulsive noise channel still achieves the space diversity and performs significantly better than DT under the same power consumption. Moreover, the proposed MAP receiver attains the lower bound derived for DF CR scheme and leads to large performance gains compared with the conventional receiving criteria, which were optimized for additive white Gaussian noise channel and memoryless impulsive noise channel.

Block-Sparse Impulsive Noise Reduction in OFDM Systems—A Novel Iterative Bayesian Approach

Using a novel block iterative Bayesian algorithm (Block-IBA), this paper presents a new impulsive noise reduction method for OFDM systems. The method utilizes the guard band null subcarriers and data subcarriers for the impulsive noise estimation and cancellation. Unlike some other general OFDM transceivers which use time-domain interleaving (TDI) to cancel impulsive noise, we design a specific receiver for bursty impulsive noise channels that removes the delay due to TDI and saves memory space. The Block-IBA first estimates the variance and the transition matrix of Markov chain model for the impulsive noise. It then iteratively estimates the amplitudes and positions of the block-sparse impulsive noise using the steepest-ascent based expectation-maximization (EM), and optimally selects the nonzero elements of the block-sparse impulsive noise by adaptive thresholding. Numerical experiments show that the proposed receiver outperforms existing receivers under the block-sparse impulsive noise environment.

Performance of Channel Coding Schemes with Interleaver under Impulsive Noise Channel

The concatenated codes or the convolutional codes which are often adopted in the communication systems are originally designed to overcome the random independent noise. As such, the performance is severely degraded over the bursty impulsive noise channels. In this paper we compare the bit error rate (BER) performance of the various different channel coding schemes with interleaver over the bursty impulsive noise channels. In particular, we focus on the existence of the bit level and/or the byte level interleavers and show from the extensive simulations that the performance can be different depending on the existence of the interleaver. From our simulations we can see that the concatenated codes with the byte interleaver between the inner and outer codes and the bit interleaver after the inner code gives the best performance among them. For the schemes which include the bit level interleaver, we also observe the interleaving gain versus the interleaver span so that it can be a good guideline to design the communication systems with the concatenated codes.

A Markov-Middleton Model for Bursty Impulsive Noise: Modeling and Receiver Design

Transmission over channels impaired by impulsive noise, such as in power substations, calls for peculiar mitigation techniques at the receiver side in order to cope with signal deterioration. For these techniques to be effective, a reliable noise model is usually required. One of the widely accepted models is the Middleton Class A, which presents the twofold advantage to be canonical (i.e., invariant of the particular physical source mechanisms) and to exhibit a simple probability density function (PDF) that only depends on three physical parameters, making this model very attractive. However, such a model fails in replicating bursty impulsive noise, where each impulse spans over several consecutive noise samples, as usually observed (e.g., in power substations). Indeed, the Middleton Class A model only deals with amplitude or envelope statistics. On the other hand, for models based on Markov chains, although they reproduce the bursty nature of impulses, the determination of the suitable number of states and the noise distribution associated with each state can be challenging. In this paper, 1) we introduce a new impulsive noise model which is, in fact, a Hidden Markov Model, whose realizations exactly follow a Middleton Class A distribution and 2) we evaluate optimum and suboptimum detections for a coded transmission impaired by the proposed noise model.