This paper presents a Joint Deep Probabilistic Subsampling framework with Cramér-Rao Bound (CRB) Integration (J-DPSC), which optimizes both transmit and receive antenna selection independently, enabling robust DoA estimation in mono-static and bi-static radar modes. Our method incorporates a differentiable worst-case CRB computation for end-to-end optimization. Additionally, sidelobe level (SLL) suppression is naturally integrated into the training process through data constraints, without requiring explicit penalty terms. Compared to prior methods, J-DPSC significantly reduces model complexity and the number of trainable parameters while maintaining high DoA estimation accuracy. Beampattern results demonstrate the effectiveness of our approach in improving performance with an acceptable SLL, making it well-suited for practical sparse array design in automotive radar applications.

High angular resolution is in high demand in automotive radar. To achieve a high azimuth resolution a large aperture antenna array is required. Although MIMO technique can be used to form larger virtual apertures, a large number of transmitter-receiver channels are needed, which is still technologically challenging and costly. To circumvent this problem, we propose a high-resolution Direction of Arrival (DoA) estimation by using multiple small radar sensors distributed on the fascia of the automobile. To exploit the diversity gain due to different target observation angles by different radars, a block Focal Under determined System Solver based approach is proposed to incoherently fuse the data from multiple small MIMO sensors. This method significantly improves the DoA estimation compared to single sensor, decreases probability of false alarm and increases probability of multiple target detection. Its performance is demonstrated through both numerical simulations and experimental results.