Abstract
Detecting high-resolution signals at binary-array receivers can yield a stochastic-resonance phenomenon. It is found, through mathematical means, that the error probability of maximum-likelihood detection (MLD) forms a convex function of the SNR; and the optimum operating-SNR increases monotonically with the signal resolution. This phenomenon encourages the use of MIMO at higher SNRs and SIMO at lower SNRs in terms of the error probability; as the former often has their signal resolutions higher than the latter. This observation also motivates a fundamental rethinking to determine whether to use MIMO or SIMO for wireless communications given binary-array receivers. In fact, there are a number of arguable advantages for SIMO, including wider coverage, higher point-to-point throughput, as well as lower complexity of the MLD. All of these are extensively investigated in this paper through both analytical work and computer simulations.