Online Learning Network Methods for a Joint Transmit Waveform and Receive Beamforming Design for a DFRC System

Abstract

Consider a joint optimal transmit waveform and receive beamforming design problem for a dual-functional radar and communication (DFRC) system. The DFRC base station sends signals to communicate with the downlink users while detecting a multiple-input multiple-output radar target. The system performance is evaluated by an affine combination between the communication multi-user interference energy and the reciprocal of the radar output signal-to-interference-plus-noise ratio. Then a joint minimization problem of the affine function is formulated, subject to constant modulus constraints. This is a typical nonconvex optimization problem. In the paper, we propose a new online learning network (OLN) scheme to solve it, by setting proper trainable network parameters, formulating a loss function, and selecting a suitable learning rate for the OLN. Simulation results are presented to demonstrate the higher performance for the DFRC system by the proposed OLN method than that by a traditional optimization method.