Hybrid Transceiver Design and Optimal Power Allocation in Downlink mmWave Hybrid MIMO Cognitive Radio Systems

Abstract

A hybrid transceiver architecture along with the optimal power allocation is conceived for a downlink millimeter wave (mmWave) multi-input multi-output (MIMO) cognitive radio (CR) system operating in the underlay mode. Towards this, the non-convex objective and constraints of the sum spectral ef-ficiency (SE) maximization problem are simplified by decoupling the hybrid precoder and combiner designs. First, considering the perfect knowledge of the downlink mmWave MIMO channel, we design the combiner at each SU. Subsequently, the front-end digital baseband (BB) precoder and analog-domain RF precoder are designed using the best-approximation problem to the capacity-optimal fully-digital precoder. Moreover, our design also considers the spatial correlation among the mmWave MIMO channels, thereby significantly reducing the computational complexity for the analog precoder/combiner design. Furthermore, in order to cancel the multiuser interference (MUI), the back-end of the BB precoder has been designed using the low-complexity zero-forcing (ZF) technique. Finally, a closed-form solution to the optimal power allocation problem is derived, which maximizes the overall SE of the downlink mm Wave MIMO CR system under the interference power constraint imposed by the primary user (PU). Our simulation findings show an improved SE compared to state-of-the-art approaches while performing close to the ideal fully-digital benchmark.