Robust MSE-Based Transceiver Optimization in MISO Downlink Cognitive Radio Network

The concurrent transmission of cognitive radios (CRs) and primary users (PUs) can achieve high spectral efficiency, provided that the interference power to each PU is limited below a certain threshold. Joint optimization of the precoding and the equalization is widely used to exploit the spatial diversity to mitigate the interference. However, its performance is sensitive to the inevitable imperfect channel state information (CSI). This paper studies the robust transceiver optimization in multiple-input and single-output downlink CR network, aiming at minimizing the worst-case per-user mean square error. The channel uncertainty model is assumed to be ellipsoidal bounded for both channel matrices and short-term channel covariance matrices. Under the uncertainty models for both cases of CSI, we first derive the strict upper bound or equivalent conditions for the constraints, then reformulate the original optimization problem into the semi-definite programming problems which can be efficiently solved. Simulation results demonstrate that the proposed schemes effectively mitigates the performance degradation due to the imperfect CSI. Their robustness with respect to the interference constraint is also validated, which is regarded as the critical part in the design of CR network.