Energy Efficiency Analysis and Optimization for Millimeter-Wave MIMO With Variable-Resolution ADCs.

In millimeter-wave (mmWave) massive multipleinput multiple-output (MIMO) systems, the surge in hardware cost and power consumption poses a huge challenge to green communication. Considering the high power consumption of the radio frequency (RF) module, we aim to design a high energy efficient uplink multi-users (MU) communication system by applying variable-resolution analog to digital converters (ADC). Firstly, a closed-form expression for the spectrum efficiency (SE) in terms of the quantization resolution of ADC is derived under hybrid beamforming (HBF) and fully digital beamforming (FDBF) architectures, respectively. We find explicit energy efficiency (EE) expressions and investigate the main factors affecting EE under two kinds of architectures. For HBF, a novel finding is that adding the number of RF chains for a fixed number of transmit data streams can effectively mitigate the quantization distortion introduced by low-resolution ADCs. Furthermore, an EE maximization problem involving joint optimization of ADC bit selection and power allocation between MU for the HBF system is formulated. Based on the concept of block coordinate ascent (BCA), we subdivide the original problem into two subproblems. The first subproblem involves power allocation, which has been reformulated as a convex problem using an equivalent quadratic transformation. The second subproblem tackles ADC bit selection, for which we propose a low-complexity search scheme that achieves performance approaching that of an exhaustive search. The simulation results demonstrate the high accuracy of the derived closed-form expression, which supports the EE analysis. The proposed EE maximization scheme has a fast convergence rate and achieves higher EE compared to existing benchmark technologies.