부산대학교 도서관

Cramér-Rao bound (CRB), as a fundamental sensing performance metric, has been widely used to optimize the radar and the integrated sensing and communication (ISAC) beamforming. Considering that the CRB requires the true value of the target parameter which cannot be obtained in practice, we develop an improved ISAC beamforming design that provides robustness against the estimation error of the target parameter. Specifically, we aim to minimize the CRB averaged over the target angle estimation error subject to a base station (BS) transmit power constraint and a downlink communication user signal-to-noise ratio (SNR) constraint. An efficient majorization-minimization (MM) based algorithm is developed to address this difficult stochastic optimization problem, where a simple closed-form solution is achieved in each iteration. Finally, simulation results validate the remarkable sensing performance gains achieved by the proposed design, especially in the challenging scenarios with low radar SNRs or high communication SNR requirements.