Cognitive radio is a potential technique to solve the spectrum shortage problem in wireless communications. Integrating wireless relaying into the cognitive radio networks can further improve the spectrum efficiency. In this study, a multiple-input multiple-output cognitive relay network is considered, where the primary network (PN) consists of one transmitter-receiver pair and the secondary network consists of multiple active source-destination pairs and non-regenerative relays. All nodes in the networks are deployed with multiple antennas. How to avoid interference to the PN is an important task in the secondary network design. The precoders of the secondary sources and the receivers of the secondary destinations are designed in a well-known zero-forcing way.

The secondary relays forward signals for the secondary source-destination pairs by beamforming. With interference cancellation constraints and individual transmit power constraints, a relay beamforming scheme is proposed to maximise the total signal-to-interference-plus-noise ratio (SINR) at the secondary destinations. Then a maximising minimum SINR relay beamforming scheme is further proposed to provide max-min fairness among the secondary source-destination pairs. The beamforming problems are formulated into the quadratically constrained quadratic fractional programming problems, and are solved by the semi-definite relaxation technique. The performances of these beamforming schemes have been verified by computer simulations.