A Cyclostationary Perspective on Noncoherent SIMO Communications

This work explores the statistical properties of the received signal in a noncoherent SIMO setting. We vectorize the classical time-space representation of the communication problem to jointly account for both domains and construct a stochastic process with a clear cyclostationary behavior. Driven by the use of massive arrays, we leverage the asymptotic structure of this class of processes and express it in the Karhunen-Loève domain. This spectral representation provides an insightful perspective on the noncoherent codeword detection problem and unveils the fundamental parameters that play a role in it.