2015IEEE AccessOpen access

Fully Parallel Turbo Equalization for Wireless Communications

Hoang Anh Ngo, Robert G. Maunder, Lajos Hanzo

Open full text 5 citations

Abstract

Iterative turbo equalization is capable of achieving impressive performance gains over the conventional non-iterative equalization having the same complexity, when communicating over channels that suffer from intersymbol interference (ISI). The state-of-the-art turbo equalizers employ the logarithmic Bahl-Cocke-Jelinek-Raviv (Log-BCJR) algorithm. However, due to the specific nature of serial data processing, the Log-BCJR algorithm introduces significant processing delays at the receiver. Therefore, in low-latency applications having a high throughput, the turbo equalizer might be deemed less attractive than its conventional counterparts. In order to circumvent this problem, in this paper, we conceived a novel fully parallel turbo equalization algorithm, which is capable of significantly reducing the data processing delay and, hence, improving both the processing latency and the attainable throughput at the receiver. The fully parallel equalizer is then combined with the fully parallel turbo decoder for improving the system performance achieved in terms of the bit error ratio. Furthermore, we propose a novel odd-even interleaver design for employment between the fully parallel equalizer and the fully parallel turbo decoder in order to reduce complexity by 50% in fully parallel turbo equalization arrangements, while retaining a comparable performance. Finally, we compare the computational complexity, latency, throughput, hardware resource requirements, and the bit error ratio of the proposed fully parallel scheme to those of a Log-BCJR-based turbo equalizer benchmarker.

Open-access reader

About this research paper

What this paper is about

Iterative turbo equalization is capable of achieving impressive performance gains over the conventional non-iterative equalization having the same complexity, when communicating over channels that suffer from intersymbol interference (ISI). The state-of-the-art turbo equalizers employ the logarithmic Bahl-Cocke-Jelinek-Raviv (Log-BCJR) algorithm. However, due to the specific nature of serial data processing, the Log-BCJR algorithm introduces significant processing delays at the receiver. Therefore, in low-latency applications having a high throughput, the turbo equalizer might be deemed less attractive than its conventional counterparts. In order to circumvent this problem, in this paper, we conceived a novel fully parallel turbo equalization algorithm, which is capable of significantly reducing the data processing delay and, hence, improving both the processing latency and the attainable throughput at the receiver. The fully parallel equalizer is then combined with the fully parallel turbo decoder for improving the system performance achieved in terms of the bit error ratio. Furthermore, we propose a novel odd-even interleaver design for employment between the fully parallel equalizer and the fully parallel turbo decoder in order to reduce complexity by 50% in fully parallel turbo equalization arrangements, while retaining a comparable performance. Finally, we compare the computational complexity, latency, throughput, hardware resource requirements, and the bit error ratio of the proposed fully parallel scheme to those of a Log-BCJR-based turbo equalizer benchmarker.

Why it matters

OpenAlex reports 5 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

Iterative turbo equalization is capable of achieving impressive performance gains over the conventional non-iterative equalization having the same complexity, when communicating over channels that suffer from intersymbol interference (ISI). The state-of-the-art turbo equalizers employ the logarithmic Bahl-Cocke-Jelinek-Raviv (Log-BCJR) algorithm. However, due to the specific nature of serial data processing, the Log-BCJR algorithm introduces significant processing delays at the receiver. Therefore, in low-latency applications having a high throughput, the turbo equalizer might be deemed less attractive than its conventional counterparts. In order to circumvent this problem, in this paper, we conceived a novel fully parallel turbo equalization algorithm, which is capable of significantly reducing the data processing delay and, hence, improving both the processing latency and the attainable throughput at the receiver. The fully parallel equalizer is then combined with the fully parallel turbo decoder for improving the system performance achieved in terms of the bit error ratio. Furthermore, we propose a novel odd-even interleaver design for employment between the fully parallel equalizer and the fully parallel turbo decoder in order to reduce complexity by 50% in fully parallel turbo equalization arrangements, while retaining a comparable performance. Finally, we compare the computational complexity, latency, throughput, hardware resource requirements, and the bit error ratio of the proposed fully parallel scheme to those of a Log-BCJR-based turbo equalizer benchmarker.

Key concepts: Turbo equalizer, Turbo, Computer science, Turbo code, Intersymbol interference, Equalization (audio), BCJR algorithm, Throughput

Related papers

Back to paper searchBrowse research topicsOriginal source
Fully Parallel Turbo Equalization for Wireless Communications — Research Paper | ScholarLens