In this paper, we consider an overlay cognitive radio network that transmits discrete-time analog sources over additive white Gaussian noise channels. Our focus is on the case where the primary and secondary sources are correlated. The secondary user (SU) knows the primary message non-causally, and allocates part of its power for transmitting the primary message. We study a hybrid digital-analog coding scheme for the SU, which is the superposition of two analog parts of the sources and a digital part of the secondary source. This coding scheme not only exploits the correlation between the sources, but also helps with the primary transmission. We derive the signal-to-noise ratio (SNR) of both PU and SU, and formulate the optimization problem as achieving maximum SNR of the SU while protecting the PU's SNR from being affected. The simulation results are shown to be consistent with our derivation.
展开▼