Evaluation and Compensation of Nonlinear Distortion in Multicarrier Communication Systems

Doctoral thesis, 2006

In modern wireless communication systems with increasing data transfer rate and mobility, designing efficient transceivers is a key issue. Among different challenges, nonlinearity of high power amplifiers is a problem that has to be overcome by means of adding extra hardware or using more efficient signal processing algorithms. Large signal envelope fluctuations of a multicarrier signal compared to a single carrier is the motivation for extensive investigations in this area. This thesis covers different methods of overcoming the distortion caused by a nonlinearity at the transmitter of a multicarrier communication system. The focus is mainly on digital baseband compensation methods for multicarrier signals undergoing a nonlinearity at the transmitter front-end. Digital baseband predistortion and nonlinear equalization are examples of methods for linearizing a nonlinear high power amplifier. Both of these methods are implemented and compared in this thesis. We also implement a method based on intercarrier interference cancellation to reduce the distortion at the receiver side by means of iterative estimation and refining the received signal. Since this method is a receiver side solution, no extra bandwidth or extra power is required, and only a more complex receiver is needed. Spectral spreading over the neighboring channel is a huge problem, that requires some means of compensation at the transmitter. The spectral spreading can be reduced by forming signals with less sensitivity to nonlinear amplifications. We implement and compare methods of reducing peak-to-average power ratio (PAPR) of a signal in order to limit the spectral outgrowth and improve the bit error rate (BER) performance of the system. Complexity, bandwidth efficiency and power efficiency are important factors that have to be considered in selection of these method. Except PAPR, other measures of the signal envelope variations can be used to evaluate the susceptibility of signal to nonlinear amplification. We use a measure related to the variance of the instantaneous power, and try to reduce that, thereby improving the performance of the system.