Asymptotic distributions and peak power analysis for uplink OFDMA signals

Statistical characterization of complex baseband signals for multicarrier multiuser systems is studied. In particular, we derive rigorously the asymptotic distribution of uplink orthogonal frequency division multiple access (OFDMA) baseband signals for different subcarrier allocation schemes. This allows us to establish some useful statistical properties of the power process of the corresponding baseband signals. We show that, in most scenarios, the power process converges to a /spl chi//sup 2/ process. This result allows us to characterize the peak-to-average power ratio (PAPR) for OFDMA signals, a parameter considered critical to real system implementation. Specifically, we are able to compare, both qualitatively and quantitatively, the PAPR for different subcarrier allocation schemes. We show first that the contiguous subcarrier allocation and equally spaced interleaving share identical PAPR. Further, we show that random interleaving, which is implemented in the existing OFDMA standard, has a larger PAPR than that of an equally spaced interleaving scheme. The complementary cumulative distribution function of PAPR for random interleaving is shown to be greater than that of equally spaced interleaving by a factor that is identical to the number of users in the OFDMA system.