Robust Linear Transceiver Design in MIMO Ad Hoc Cognitive Radio Networks with Imperfect Channel State Information

The joint linear transceiver design for Multiple-Input Multiple-Output (MIMO) \adhoc cognitive radio networks when the channel state information (CSI) is uncertain is discussed in this paper. The uncertainty in CSI is modeled using Stochastic Error (SE) and Norm Bounded Error (NBE) models. The Sum-Mean Square Error (SMSE) performance is used to formulate the design problem. To optimize the network, the transmit power budget for secondary user transmitters and the maximum allowed interference at primary user receivers are constrained. In the design methodology, the network parameters are optimized to best serve the users when the worst possible channel realizations occur. It is shown that for the SE model of uncertainty, this problem can be cast as a Second Order Cone Problem (SOCP), while for the NBE model, the problem becomes a Semi-Definite Program (SDP). These two problems are solved efficiently using the numerical solver packages YALMIP and SDPT3. Finally simulation results are presented to evaluate the performance of the proposed methods.