A Fast MUSIC Algorithm for High-Quality Real-Time Point Clouds Acquisition

To enhance radar angular resolution while meeting real-time processing requirements, a fast Multiple Signal Classification (MUSIC) algorithm based on dimensionality reduction is proposed in this paper. By analyzing the configuration of Frequency Modulated Continuous Wave (FMCW) radar's antenna array and its echo signal models, we decompose the complex joint angle estimation into two simpler one-dimensional estimations. Simulation experiments demonstrate that the proposed MUSIC algorithm achieves a 100-fold speed improvement with no significant accuracy loss. When applied to real-time point cloud acquisition, the proposed algorithm outperforms the conventional Three-Dimensional Fast Fourier Transform(3D-FFT) and Texas Instruments (TI) method by 82.9% and 36.9%, respectively, while achieving better point cloud separation results.