Nonlinear Equalization for a Frame-Differential IR-UWB Receiver

Project Type: Master/Diploma Thesis
Student: Krall Christoph

  For a frame-differential autocorrelation receiver structure, typically used in Ultra-wide-band (UWB) technology, equalization is investigated. The radio channel, due to its multipath nature, distorts the received signal. Therefore the received data suffers from inter-symbol interference (ISI) at high data-rates. This ISI should be compensated by an equalizer. The channel's impulse response is assumed to be constant for a limited time. The design of the equalizer should also consider noise.

This thesis provides an overview of different methods for equalizing the nonlinear effects. Furthermore, techniques which consider the noise, which is enhanced by the receiver, are developed.

It is shown that the Volterra system describing the signal model doesn't possess a minimum-phase linear part. Therefore, perfect equalization is not possible. Minimum mean square error (MMSE) techniques are used to achieve linear equalization while minimizing the effects of the noise. Further improvement of the performance in equalization is done via oversampling the receiver front-end, which leads to fractionally spaced methods for equalization.

 

It is shown that the identification of a nonlinear system is not possible with a BPSK signal. Nonlinear adaptive approaches with the use of adaptive polynomial filters are shown. The equalizer techniques applied to the back-end of the receiver structure increase the performance significantly.