Network coding versus superposition coding for two-way wireless communications

Wireless network coding has attracted significant attention recently over relay channels. In this paper, we are interested in two-way wireless communication and explore the comparative advantages of two popular network coding strategies, binary network coding (NetC) and superposition coding (SupC), also named as analog network coding. In the literature, most works focused on NetC under a three-step framework and recently also SupC but under a two-step one. However, there is still little understanding on the truly optimal configuration. In addition, most works assumed channel knowledge at the transmitters while ignoring the presence of the direct link, which may not be valid in a fading environment. In this work, we focus on the outage performance instead and take into consideration also the time-varying nature of the direct link. Outage regions of various schemes are theoretically characterized and their optimal operating conditions are identified, from which the best combined strategies are derived in terms of the maximum goodput and robustness to channel knowledge defects. Interesting new findings together with insightful discussions are provided. ©2009 IEEE