Correcting Deviations of Emulated Spatial Correlation for Prefaded Signal Synthesis Method With Limited Number of Random Drops in MIMO OTA Testing

The prefaded signals synthesis (PFS) is a well-established over-the-air (OTA) testing solution, which can reconstruct the desired spatial profiles of the target channel by adjusting the power weights of the multiple probes. However, practical channel emulation campaigns tend to use a limited number of random drops due to limited time budget and hardware capacity constraint, which may cause the emulated spatial correlation to be deviated from the expected one due to the loss of ergodicity. In this article, the properties of the emulated channel under the PFS method are reexamined for a given set of initial subpath phases that are randomly configured at different probes. The analytical expressions of the spatial–temporal correlation function (STCF) are thus derived by averaging over time, compared with the general manipulations of ensemble averaging in the literature. Our derivations indicate that the spatial correlation function (SCF) of the emulated channel may deviate from the expected form for a given set of initial subpath phases in the classic PFS method. An improved configuration method for initial subpath phases is proposed. Simulation results show that the proposed method can reproduce the exact SCF in any random drop while unaffecting other important channel properties, for example, the temporal correlation function (TCF).