Estimation Based Feedback Reduction for Online Fountain Codes

Abstract

Online fountain codes attract much attentions for their good intermediate performance. However, the feedback requirement limits the application scenario of the coding scheme. In this paper, we reduce the feedback requirement of online fountain codes with no build-up phase (OFCNB). First we analyze the relationship between the overhead and the recovery rate as well as the relationship between the overhead and the butter occupancy. Based on the analysis, we proposed an estimation based degree selection (EBDS) scheme for OFCNB, where the coded symbol degree is selected based on the estimation of decoding state rather than through feedback. Simulation results validate the effectiveness of the proposed analysis. Compared with other feedback reduction scheme, the proposed OFCNBEBDS scheme requires less full recovery overhead with similar number of feedback transmissions.