Research Article
Two Degree Forest Based LT Codes with Feedback
@INPROCEEDINGS{10.1007/978-3-319-66625-9_23, author={Liang Liu and Feng Liu}, title={Two Degree Forest Based LT Codes with Feedback}, proceedings={Communications and Networking. 11th EAI International Conference, ChinaCom 2016, Chongqing, China, September 24-26, 2016, Proceedings, Part I}, proceedings_a={CHINACOM}, year={2017}, month={10}, keywords={LT codes Two degree forest Feedback channel Robust soliton distribution}, doi={10.1007/978-3-319-66625-9_23} }- Liang Liu
Feng Liu
Year: 2017
Two Degree Forest Based LT Codes with Feedback
CHINACOM
Springer
DOI: 10.1007/978-3-319-66625-9_23
Abstract
The performance of belief propagation (BP) decoding algorithm for LT codes is significantly deteriorated, as the data-block length decreases, since the randomly encoding of LT codes causes lots of wasted output symbols, which is helpless for decoding. To solve this problem, this paper provides two degree forest based LT codes in order to help the sender to send badly needed symbols to accelerate decoding throughout entire receiving process. Through gathering two degree output symbols into separable trees, the decoder can easily get and feedback the indexes of the badly needed input symbols. Simulation results show that, in the short data-block length case, two degree forest based LT codes achieve much lower coding overhead, consume much smaller storage resources, and need less feedback opportunities compared with current LT codes algorithms.