Fast Randomized-MUSIC for Mm-Wave Massive MIMO Radars

Subspace methods are essential to high-resolution environment sensing in the emerging unmanned systems, if further combined with millimeter-wave (mm-Wave) massive multi-input multi-output (MIMO) technique. The estimation of signal/noise subspace, as one critical step, is yet computationally complex and presents a particular challenge when developing low-complexity automotive radars. In this work, we develop fast randomized-MUSIC (R-MUSIC) algorithm, which exploits random matrix sketching estimate signal subspace via approximated computation. Our new approach substantially reduces time complexity acquiring high-quality subspace. Moreover, accuracy R-MUSIC suffers no degradation unlike others counterparts, i.e. angle arrival (AoA) attained. Numerical simulations provided validate performance our method. As shown, it resolves long-standing contradiction MIMO radar processing, hence have great potentials real-time super-resolution sensing.