Rate-Adaptive Probabilistic Shaping Using Polar Codes
Shajeel Iqbal
Technical University of Denmark
Polar codes asymptotically achieve the capacity of discrete memoryless symmetric channels with low complexity encoding and decoding algorithms that outperform, e.g., WiMAX LDPC codes. Furthermore, the successive cancellation decoder for polar codes does not display an error floor, in contrast to LDPC codes that display an error floor if they are not carefully designed. Hardware based polar decoders of short code lengths are already under investigation for 5G due to their ability to achieve Gbps throughput with less area, power and energy consumption than WiMAX LDPC codes. As a potential candidate for future lightwave systems, we study polar codes using list decoding. Punctured polar codes are a viable solution to the rate-adaptivity problem at the expense of performance degradation using quasi-uniform puncturing. We demonstrate how many-to-one probabilistic shaping can recover the performance loss related to the puncturing of polar codes. We then jointly optimize the many-to-one shaper with the polar-coded frozen bits, thus achieving additional shaping gain w.r.t. standard rate-adaptive punctured polar-coded systems.
Shajeel received his Bachelor in Electronics Engineering from National University of Sciences and Technology, Pakistan in 2012. He received his Master in Information and Communication Engineering from Chosun University, South Korea in 2016. He is currently pursuing his PhD studies at the Technical University of Denmark under the supervision of Søren Forchhammer. Shajeel is visiting Technical University of Munich from Oct 2018-Jan 2019 as a part of his PhD studies. His research interests include coding theory and information theory for optical communication systems.