Font Size: a A A

Turbo decoding for packet data systems

Posted on:1999-03-24Degree:Ph.DType:Dissertation
University:Georgia Institute of TechnologyCandidate:Narayanan, Krishna RamaFull Text:PDF
GTID:1468390014468897Subject:Engineering
Abstract/Summary:
Parallel concatenation of convolutional codes used in conjunction with iterative decoding, known as Turbo codes, has proven to be a remarkable development in error control coding since its introduction in 1993. More recently, serial concatenation of interleaved codes (SCCC) has also been shown to provide excellent coding gains. Iterative decoding of interleaved convolutional codes is considered by many to be the solution to achieving reliable communication through noisy channels at signal-to-noise ratios (SNR) close to the theoretical Shannon limit. There are a lot of applications including deep space communications, satellite communications and magnetic recording, that can potentially benefit from these advances in coding theory. One such an application is the transmission of packetized data over time-varying channels. In this study, some methods are presented to improve the performance of Turbo codes when used for transmitting packetized data over non-stationary Gaussian channels. The application of SCCC to packet data systems that require very low bit error rates is also investigated.;A bandwidth efficient serial concatenation scheme with a Bose-Chaudhuri and Hocquenghem (BCH) outer code and Turbo inner code is proposed to reduce the error floor of Turbo codes. List decoding of Turbo codes is proposed as a technique to improve the frame error rate performance of Turbo codes at high SNRs. Asymptotic gains are derived for maximum likelihood (ML) list decoding of Turbo codes on AWGN and flat Rayleigh fading channels. Then, two practical list decoders are proposed to generate a list of paths at the Turbo decoder output. Some retransmission schemes based on packet combining techniques are proposed to improve the throughput efficiency of the system at low SNRs. A receiver structure that combines a list decoder and a retransmission scheme with the iterative Turbo decoder is proposed. Then, iterative demodulation and decoding is treated as a special case of SCCC and, performance bounds are derived. Finally, trellis coded interleaved continuous phase modulation (CPM) is analyzed from a serial concatenation perspective. New performance bounds are derived for full response CPM on AWGN and fading channels by explicitly considering the effect of the interleaver on the distance spectrum.
Keywords/Search Tags:Turbo, Decoding, Data, Channels, Packet, Iterative, Concatenation
Related items