Abstract
Hybrid antenna arrays are considered as a promising tool for future wireless networks that allows enhanced positioning accuracy with limited number of RF chains. In this paper, we develop an optimization algorithm for analog elements connecting the array elements and the RF chains based on an iterative element-by-element adjustment with controllable convergence behavior for different hybrid architectures. Our algorithm relies on electromagnetic modeling accounting for near/far field and array geometry effects with mutual coupling correction. Assuming random element selection for updates, we apply Markov chain modeling for semi-analytical (analytical for the given scenario realization) investigation of partially connected switched configurations and demonstrate a trade-off between the stationary localization accuracy and the convergence rate. We extend our Markov chain analysis to a hybrid network of spatially distributed sub-systems and show a possibility of significant localization accuracy improvement with joint non-coherent sub-system processing. For higher dimension fully connected architectures, we verify the semi-analytic results by means of simulations with electromagnetic modeling.