Abstract
For the non-regenerative multiple-input multiple-output (MIMO) cognitive multi-relay network, we propose an optimal relay selection and beamforming scheme which maximizes the capacity of the secondary user by selecting the best cognitive MIMO relay subject to the transmit power constraints at the relays and the interference power constraints at the primary users. In our proposed scheme, the relay selection and beamforming optimization problem is solved separately by employing convex semidefinite programming (SDP) through rank-one relaxation and Charnes-Cooper transformation. It is shown that our proposed scheme achieves higher average capacity than the convectional relay beamforming schemes and relay selection schemes.