Spectrally Compatible Waveform Design for Large-Scale MIMO Radar Beampattern Synthesis With One-Bit DACs

Abstract

Multi-input multi-output (MIMO) radar with massive antennas is promising for high resolution applications. However, a big challenge of this system is that the hardware cost and power consumption will increase significantly, if high-resolution quantizers are adopted. In this article, we consider MIMO radar deployed with one-bit digital-to-analog converters, and investigate the problem of designing one-bit transmit sequence with good spatial and spectral properties. Specifically, the one-bit waveform design problem is formulated by minimizing the mean-square error between the desired and designed transmit beampatterns, subject to spectral constraints. The resulting problem, including a nonconvex quartic objective and a nonconvex discrete constraint, is NP-hard, and an alternating optimization (AltOpt) framework with the aid of “almost equivalent” criterion is thereby developed to handle it. Particularly, in the AltOpt framework, a low-complexity algorithm is developed based on the alternating direction method of multipliers approach. Numerical simulations are provided to show the advantages of the proposed method over the state-of-the-art techniques in terms of the spatial and spectral properties as well as computational complexity.