Paper Info
Title
RT-Polar: An HARQ Scheme with Universally Competitive Rates
Abstract
We present a construction for a universal channel code with feedback using Polar Codes. Our construction includes an error detection mechanism that is used to compute the ACK/NACK feedback directly from the received vector, without a higher layer CRC. Our scheme, termed the Repeat-Top Polar Code (RT-Polar), builds on a rate-compatible Polar Code and retransmits the t message bits sent over the most reliable polarized good channels over the least reliable good channels. At the decoder, these two t-bit strings are decoded and compared to detect an error. Through simulations, we illustrate the universal performance of our scheme for a binary symmetric channel with an unknown flipover probability. Our scheme performs comparably with a genie-aided scheme, where the detection mechanism is assumed to be error-free, for practically relevant message lengths of roughly 512 bits; this is the first instance of such a universal performance reported in literature. The proposed scheme is suitable for use as a HARQ in low-latency communication where including a higher-layer CRC will induce computational delays.
Year
DOI
Venue
2018
10.1109/ITW.2018.8613460
2018 IEEE Information Theory Workshop (ITW)
Keywords
Field
DocType
HARQ scheme,universally competitive rates,universal channel code,error detection mechanism,ACK/NACK feedback,received vector,higher layer CRC,Repeat-Top Polar Code,RT-Polar,rate-compatible Polar Code,message bits,reliable polarized good channels,decoder,t-bit strings,universal performance,binary symmetric channel,unknown flipover probability,genie-aided scheme,error-free,practically relevant message lengths,word length 512.0 bit
Hybrid automatic repeat request,Binary symmetric channel,Computer science,Communication channel,Algorithm,Error detection and correction,Theoretical computer science,Redundancy (engineering),Polar code,Decoding methods,Benchmark (computing)
Conference
ISSN
ISBN
Citations 
2475-420X
978-1-5386-3600-8
0
PageRank 
References 
Authors
0.34
0
2
Name
Order
Citations
PageRank
Soumya Jyoti Banerjee153.95
Himanshu Tyagi211813.82