Abstract
Reconfigurable intelligent surface (RIS) has been introduced into radar systems to generate an extra path from the transceiver to the target by artificially reflecting radar signals, and thus provides additional degrees-of-freedom to enhance the performance. This paper studies the beampattern synthesis for a RIS-aided radar by minimizing the peak sidelobe level and the mismatch error between the mainlobe beampattern and the desired one. The proposed beampattern synthesis is described as a nonconvex constrained optimization problem to jointly design the digital beamforming of the radar transmitter and the passive beamforming of the RIS. To solve the problem, the auxiliary variables are first introduced to simplify the nonconvex constraints. Then all variables are alternatively optimized by using the alternating direction method of multipliers (ADMM), in which the active beamforming is provided by its close-form feasible solution and the passive beamforming is solved by the Riemannian manifold optimization. Numerical results show that the proposed RIS-aided radar can supply a significantly lower sidelobe level than typical monostatic radars.