In post-selection hypothesis testing, a third outcome is added, corresponding to not selecting any of the hypotheses. In this paradigm, minimum error for various cases is characterized in literature conditioned on the fact that one of the selected outcomes occurs. However, not selecting any of the hypotheses decimates the quality of testing. There is no discussion on the quality of post-selection hypothesis testing. In this work, we characterize the quality of post-selection hypothesis testing by defining a new metric acceptance. We find out measurements which lead to optimal post-selection error. For these measurements, we have derived the expression for acceptance in post-selection symmetric and asymmetric hypothesis testing problems. We have also given closed-form expressions in some specific cases.

In mmWave networks, a large or nearby object can obstruct multiple communication links, which results in spatial correlation in the blocking probability between a user and two or more base stations (BSs). This paper characterizes this blocking correlation and derives its impact on the signal-to-interference-plus-noise ratio (SINR) of a mmWave cellular network. We first present an exact analysis of a 1D network and highlight the impact of blocking correlation in the derived expressions. Gaining insights from the 1D analysis, we develop an analytical framework for a 2D network where we characterize the sum interference at the user by considering the correlation between the blocking of serving and interfering links. Using this, we derive the SINR coverage probability. Via simulations, we demonstrate that including blockage correlation in the analysis is required for accurate characterization of the system performance, in particular when the blocking objects tend to be large.

Past literature on the analysis of mmWave cellular network's performance in the presence of random blockages has assumed independent blocking of each link. In reality, a large blockage can block multiple links resulting in a correlation in blocking of these links. This paper focuses on deriving the impact of this correlation on the SINR performance of a mmWave cellular system. In particular, we consider the conditional probability of having line-of-sight (LoS) and non-LoS (NLoS) interferers conditioned on serving link being LOS or NLoS. We derive the distribution of the serving base-station (BS) distance from a typical user. We characterize the sum interference at the user by considering the correlation between blockage of serving and interfering links and derive the SINR coverage probability. Via numerical simulations, we show that including blockage correlation in the analysis is required in accurate characterization of the system performance, in particular when the blockages tend to be large.