Characterizing the Performance of Wireless Communication Architectures via Basic Diophantine Approximation Bounds

Consider a wireless network where several users are transmitting simultaneously. Each receiver observes a linear combination of the transmitted signals, corrupted by random noise, and attempts to recover the codewords sent by one or more of the users. Within the context of network information theory, it is of interest to determine the maximum possible data rates as well as efficient strategies that operate at these rates. One promising recent direction has shown that if the users utilize a lattice-based strategy, then a receiver can recover an integer-linear combination of the codewords at a rate that depends on how well the real-valued channel gains can be approximated by integers. In other words, the performance of this lattice-based strategy is closely linked to a basic question in Diophantine approximation. This chapter provides an overview of the key findings in this emerging area, starting from first principles, and expanding towards state-of-the art results and open questions, so that it is accessible to researchers with either an information theory or Diophantine approximation background.