Complexity Reduction in Beamforming of Uniform Array Antennas for MIMO Radars

Covariance matrix design and beamforming in multiple-input multiple-output (MIMO) radar systems have always been a time-consuming task with a substantial number of unknown variables in the optimization problem to be solved. Based on the radar and target conditions, beamforming can be a dynamic process and in real-time scenarios, it is critical to have a fast beamforming. In this paper, we propose a beampattern matching design technique that is much faster compared to the well-known traditional semidefinite quadratic programming (SQP) counterpart. We show how to calculate the covariance matrix of the probing transmitted signal to obtain the MIMO radar desired beampattern, using a facilitator library. While the proposed technique inherently satisfies the required practical constraints in covariance matrix design, it significantly reduces the number of unknown variables used in the minimum square error (MSE) optimization problem, and therefore reduces the computational complexity considerably. Simulation results show the superiority of the proposed technique in terms of complexity and speed, compared with existing methods. This superiority is enhanced by increasing the number of antennas.