Impact of sampling rate on PSD of OFDM signal with SSOP sidelobe suppression

We recently proposed an efficient "sidelobe suppression with orthogonal projection" (SSOP) scheme for reduction of out-of-band emission of OFDM signals. The SSOP scheme uses a preceding matrix to generate zeros at prescribed frequencies in the sidelobe region and it is numerically observed that its suppression effect is affected by the sampling rate. In this paper, we analytically characterize this effect by considering analog and digital implementations of the SSOP scheme, which correspond to a zero sampling rate and an IFFT implementation with non-zero sampling rates, respectively. Both implementations are modeled and analytical expressions for the corresponding spectra are developed. It is shown that the analytical results match the simulation results very well and the suppression performance is generally proportional to the sampling rate. The analog SSOP-OFDM has true zeros at prescribed frequencies and thus results in significant reduction of out-of-band emissions. The digital implementation does not result in true zeros in the spectrum but nevertheless results in significant sidelobe reduction if the frequency resolution associated with the IFFT is sufficiently high.