Blind Index Coding over Wireless Channels: The Value of Repetition Coding

The majority of modern wireless communication systems rely on transmission of discrete information packets. An immense amount of work has been dedicated to studying such wireless packet networks. In fact, understanding of some scenarios is comprehensive from both theoretical and practical perspectives (e.g., ALOHA [1], 802.11 [2], etc.). Unfortunately, among the poorly understood scenarios are those expected to be most prominent in next-generation wireless systems: multiuser communication over multi-hop relay networks. Ostensibly, principles from (wireline) network coding should be valuable in such scenarios, and indeed codes based on wireline random linear network codes can achieve capacity for certain single unicast and single multicast problem settings [?]. However, such schemes are provably suboptimal when multiple messages must be communicated, and even the capacity of a simple two-hop broadcast channel remains an open problem [?].