Conditions for Capacity-Achievability of Polar IR-HARQ Schemes

Abstract

Researchers have been trying to develop polar IR-HARQ schemes in the past years, and incremental freezing is introduced as a scalable IR-HARQ scheme. However, its finite-length block performance is not ideal and sometimes is even worse than Chase combining. We had a key observation that there should be a final combining stage for all transmissions, with which the polar IR-HARQ performance could be significantly improved in the finite block-length regime. In this paper, we discuss models based on this observation and show the conditions for them to achieve the capacity of the underlying channels.