Joint velocity and angle estimation for single-cycle TDM MIMO automotive radar

Time -division multiplexing (TDM) multiple -input multiple -output (MIMO) radar systems have gained popularity in automotive applications, thanks to their high -resolution and cost-effective properties. However, when the relative motion between the target and the radar is significant and possibly fast time -varying, TDM MIMO systems suffer for reduced maximum unambiguous velocity and angle -Doppler coupling. In this work, a novel off -grid estimation method for TDM MIMO systems is developed that operates on a single cycle and is able to accurately estimate unambiguously the instantaneous velocity and the target direction of arrival (DOA). We model the angle -Doppler coupling through a space-time steering vector. Then, an iterative adaptive approach (IAA) -based method is exploited to reconstruct the 2D velocity -angle spectrum. Additionally, we account for offgrid errors when the targets do not lie on the search grids. A coarse -to -fine strategy is proposed to alleviate the computational burden, where a maximum likelihood (ML) estimator iteratively corrects the off -grid velocity and DOA estimates of each target. Simulated and experimental data are processed to demonstrate the effectiveness of the proposed method.