BER Reduction Using Precoded OFDM in Power-line Channel

Sudeepa Madushanka Herath, Regina Reine, Zhuquan Zang, Filbert H. Juwono

Research output: Contribution to journalConference articlepeer-review

Abstract

Power-line Communications (PLC) effectively uses the pervasive power grid as a high data rate communication medium. Despite its advantages, the presence of severe multipath fading and susceptibility to impulsive noise make reliable communication in a PLC system challenging. In PLC systems, orthogonal frequency-division multiplexing (OFDM) is an attractive choice mainly due to its resilience to multipath fading. However, OFDM signals suffers from large peak-to-average power ratio (PAPR) thus resulting in non-linear signal distortions at the transmitter. Therefore, for an OFDM-based PLC system, the combined effect of large PAPR and the presence of impulsive noise results in high bit error rate (BER). This paper proposes the use of precoding at the transmitter to reduce PAPR, and blanking non-linearity at the receiver to mitigate impulsive noise. Precoding is a data-independent and robust PAPR reduction scheme that essentially distributes the power of the individual OFDM symbols over the entire OFDM block. Blanking is an effective impulsive noise mitigation technique that results in better BER performance at the receiver. Simulation results show that, with the combination of precoding and blanking, significant gains in BER performance for an OFDM based PLC system can be obtained with minimal increase of transceiver complexity.

Original languageEnglish
Article number012040
JournalIOP Conference Series: Materials Science and Engineering
Volume495
Issue number1
DOIs
Publication statusPublished - 2019
Externally publishedYes
Event11th Curtin University Technology, Science and Engineering International Conference, CUTSE 2018 - Miri, Sarawak, Malaysia
Duration: 26 Nov 201828 Nov 2018

Fingerprint

Dive into the research topics of 'BER Reduction Using Precoded OFDM in Power-line Channel'. Together they form a unique fingerprint.

Cite this