Arithmetic-BICM for Seamless Rate Adaptation for Wireless Communication Systems

Abstract

Bit-interleaved coded modulation (BICM) has been widely used in capacity-achieving communication systems. In this paper, we propose an enhanced BICM, namely, arithmetic-BICM (A-BICM), which is able to achieve high channel throughput and seamless rate adaptation in time-varying wireless channels. The A-BICM generates modulation symbols through arithmetic weighted sum operations, and it adapts to data rate via adjusting the number of transmitted symbols. Our contributions in the design of A-BICM are twofold. First, we use an extrinsic information transfer chart to analyze a demapping procedure, and this facilitates the selection of the weighting multiset, which is a key design parameter in A-BICM mapping. Second, we employ a systematic channel code in A-BICM and design an adaptation scheme to strike a balance between the performances in both high- and low-SNR regions. Simulation results show that the A-BICM outperforms the conventional BICM with length-864 low-density parity check coding in both additive white Gaussian noise and fading channels. The experiments in real WLAN environments reveal that A-BICM achieves 42.1% and 20.9% throughput gain over conventional BICM and hybrid automatic repeat request rate adaptation schemes. Furthermore, the channel coding rate and its constellation in A-BICM do not need to change within a fairly wide SNR region from 1 to 30 dB to reduce communication overhead.