Font Size: a A A

Peak-to-average power ratio reduction in OFDM systems

Posted on:2009-03-14Degree:Ph.DType:Thesis
University:University of Alberta (Canada)Candidate:Wang, LuqingFull Text:PDF
GTID:2448390005950241Subject:Engineering
Abstract/Summary:
Orthogonal Frequency Division Multiplexing (OFDM) suffers from high Peak-to-Average Power Ratio (PAR) issues. A large PAR requires a linear High Power Amplifier (HPA), which, however, is inefficiently used. Moreover, the combination of an HPA having insufficient linear-range and a large PAR leads to in-band distortion and out-of-band radiation. Various PAR reduction techniques have been proposed.;This thesis focuses mainly on the PAR reduction of OFDM systems by using tone-reservation, sign-selection, and coding techniques. We analyzed the clipping noise by approximating it as a series of parabolic pulses. Our analysis explains peak regrowth and the constant clipping noise power spectrum over the whole OFDM band. We also establish the roughly proportional relationship between the clipping noise at the end of several clipping and filtering iterations, and that generated in the first iteration. The constant of proportionality is estimated via the level-crossing theory. Two algorithms are proposed to reduce the PAR under the tone-reservation constraints. These algorithms scale the filtered first-iteration clipping noise by a constant or adaptively-calculated factor to compensate for peaks above the threshold. The complexity analysis and simulation show that our proposed algorithms achieve a larger PAR reduction and lower complexity than the active-set algorithm.;The PAR can also be reduced by optimizing the signs of data symbols. The optimal signs may be transmitted to the receiver as side information to correctly decode data symbols. In this thesis, we propose an adaptive mapping scheme to eliminate the need for side information at the receiver. We also propose several algorithms (based on using a stochastic search or clipping noise as a guide) to solve the associated discrete optimization problem of the sign-selection technique. The complexity analysis and simulation confirm the complexity advantages of the proposed algorithms compared to the selective mapping and the derandomization algorithms. Finally, we generalize the Rudin-Shapiro sequence. Constructed from an initial Phase Shift Keying (PSK) sequence, the generalized sequence increases the coding rate at the cost of an increased PAR. By optimizing the signs of the initial sequence, the PAR of the generalized sequence can be further reduced.
Keywords/Search Tags:PAR, OFDM, Power, Clipping noise, Sequence
Related items