Performance analysis and impairment mitigation of digital subscriber lines in the face of impulsive noise

Abstract

Digital subscriber lines (DSL) continue to evolve in supporting home networks, whilst relying on a hybrid architecture of twisted pairs and optical fiber, as a benefit of its cost-efficiency and high quality of service. However, the preponderance of impulsive noise stifles the further improvement of the system performance. Explicitly, the throughput, reliability as well as delay, have to satisfy a range of stringent communication requirements. Against this background, in this thesis, we analyze the performance of DSL system in the face of impulsive noise. Moreover, we propose a joint impulsive noise estimation and data detection algorithm for alleviating the detriments caused by impulsive noise, which are briefly elaborated on below.

We commence by reviewing the state-of-the-art in impulsive noise mitigation, both at the transmitter as well as at the receiver. These encouraging works also consolidate our motivation of this thesis. Moreover, a range of metrics suitable for characterizing the above-mentioned deleterious is highlighted, followed by a survey of both empirical and simplified mathematical noise models.

Secondly, we analyze the bit-error ratio (BER) performance of DSL systems in the face of impulsive noise. Explicitly, we consider multi-carrier systems in the face of hidden semi-Markov model based impulsive noise. Moreover, since the overall BER performance of the system is dependent both on the occurrence probability of impulsive noise and on the specific distribution of the impulsive noise’s amplitude, we analyze the statistics of impulsive noise samples both in the time- and in the frequency-domain, and then investigate the amplitude distribution of impulsive noise after multi-carrier demodulation with the aid of the Central Limit Theorem. Based on the above results, closed-form BER formulas are derived. Our extensive simulation results verify the accuracy of the analysis in quantifying the deleterious influence of impulsive noise.

Furthermore, given that hybrid automatic repeat request (HARQ) constitute a popular technique of eliminating the effects of impulsive noise, we investigate the performance of the Chase combining aided HARQ and of plain HARQ schemes in a low-density parity-check (LDPC) coded multi-carrier system inflicting impulsive noise, in terms of its outage probability and the average number of transmission attempts as well as goodput. Specifically, a modified density evolution technique is proposed, which is then involved for characterizing the outage probability performance of the system in a finite block-length regime. The accuracy of our analysis is verified by extensive simulation results, which confirms our expectation that Chase-combining HARQ outperforms the plain HARQ.

Additionally, we conceive a joint impulsive noise estimation and data detection algorithm for LDPC-coded multi-carrier systems. More explicitly, first of all, we propose a semi-blind estimation method, which is capable of estimating the arrival of noise impulses inferred by evaluating the power of impulsive noise with an adequate accuracy. Secondly, in order to improve the accuracy of impulsive noise estimation, we propose a decision-directed (DD) method for the second stage of channel decoding and data detection with the aid of extrinsic information transfer (EXIT) charts. Our proposed two-stage scheme is capable of approaching the performance of the idealistic scenario of perfectly knowing both the arrival time and the instantaneous power of impulsive noise. Moreover, we analyze the mean square error (MSE) of the proposed schemes in order to quantify the estimation accuracy and to reduce the estimation complexity. Our simulation results demonstrate that our proposed scheme is capable of achieving a near-capacity performance in the LDPC-coded multi-carrier DSL system in the presence of impulsive noise.

More information

Identifiers

ORCID for Lajos Hanzo: orcid.org/0000-0002-2636-5214

Catalogue record

Export record