Robust Beamforming for Integrated Sensing and Communication with Angle Uncertainty

This work examines robust beamforming for integrated sensing and communication (ISAC) systems. We consider a vehicle-to-infrastructure (V2I) downlink system with a dual-functional radar-communication (DFRC) setup, where the communication signal sent by the roadside unit to the target vehicle is also used for sensing of the vehicle's angular position. The angle information is then utilized to further improve the communication in the next time slot. We aim to design transmit signals that can simultaneously achieve communication rate and sensing requirements under angle uncertainty. The transmit precoder for each time slot is optimized using angle estimates from the previous time slot, incorporating prediction errors into the performance metrics. The proposed robust beamforming design minimizes power consumption while meeting constraints on the achievable rate and the Bayesian Cramér-Rao bound (BCRB) of the angle estimate, for all potential angular positions. We employ the region of interest (ROI) approach to incorporate angle uncertainty and adopt semidefinite relaxation (SDR) techniques to efficiently solve the optimization problem. Simulation results show the effectiveness and robustness of the proposed algorithm in the presence of prediction errors.