| This thesis investigates statistical methods of compensating for nonlinear distortions in the Orthogonal Frequency Division Multiplexing (OFDM) transmission path without memory. In accordance with accepted methodology, because of limitations of digital signal processors the compensation methods are applied at the baseband, though the actual effects are in the Intermediate Frequency (IF) and Radio Frequency (RF) stages. The thesis first introduces an effective algorithm to obtain the inverse function of any invertible memoryless nonlinear system based on the cumulative distribution functions (CDFs) of the system input and output signals. Using this function, a low-complexity predistorter at the transmitter is proposed to compensate for nonlinearity in the high-power amplifier (HPA), so that the overall system has a linear transfer function. To demonstrate performance improvements of the proposed predistortion algorithms, bit error rate (BER) performance, spectral spreading and constellation warping are analyzed.; This thesis also proposes an efficient technique to compensate for memoryless nonlinear channel distortions at the receiver with the objective of providing a linear equivalent channel. (Abstract shortened by UMI.)... |
