“…The preprocessing was conducted using MATLAB R2020b on a computer with Intel i7-7700K CPU to detect potential targets, cropped the range-azimuth-elevation cubes from the generated 3D radar imaging. During preprocessing, the used hyper-parameters are summarized here: probability of false alarm in CFAR (1 × 10 −4 ), Range FFT points (256), Velocity FFT points (64), Angle FFT points for azimuth (86), Angle FFT points for elevation (16), clustering threshold ([ε r , ε v , ε a ] = [10,8,8]), cropped cube size along range, azimuth and elevation (24,24,10), and cube amplitude normalization value (1 × 10 5 ). The cropped and normalized cubes are stored at the local PC for following training and testing usage.…”