Convolutional coding schemes with convenient power spectral density characteristics
2014-09-30T08:00:08Z (GMT) by
This thesis analyses the effects caused by the introduction of convolutional coding schemes over the power spectral density (PSD) of impulse radio (IR) based ultra wideband (UWB) communication systems. This is an important topic since UWB emissions are expected to comply with spectral masks imposed by various regulatory bodies. The spectral analysis of convolutionally codedlMarkov-driven IR-based UWB signals is carried out by assuming that the data stream at the encoder input is generated by a binary Markov source (BMS). The signal model covers a wide variety of IR-based modulation schemes such as pulse position modulation (PPM), binary phase shift keying (BPSK), pulse amplitude modulation (P AM), biorthogonal PPM (BOPPM), PAMlPPM and pulse shape modulation (PSM), combined with periodic or random time hopping (TH) and/or direct sequence (DS) multiplication. In addition the inclusion of attenuation and random jitter has been accounted for in the signal model. Novel closed form PSD expressions for convolutionally codedlMarkov-driven IR-based UWB signals are obtained by performing the spectral analysis of the signal model. The formulas provided clearly identifY. the relative contributions of different parameters allowing enhanced design of UWB systems. It is shown that convolutional encoders can be used to generate IR-based UWB signals with advantageous PSD characteristics. The analysis presented in this dissertation focuses on spectral line suppression/elimination. Novel maximum free distance (MFD) binary to M-ary convolutional encoders with spectral line suppression capabilities for M-ary PPM based UWB systems are presented. Moreover, novel MFD binary convolutional encoders which generate BPSKlQ-BOPPM IR-based UWB signals whose PSD is spectral line-free are introduced. Results show that for IRbased UWB systems these convolutional encoders offer improved PSD characteristics when compared to the standard non-coded case and the best convolutional encoders known. As well, it is demonstrated that by using these convolutional encoders improved PSD characteristics and improved bit error rates (BER) can be achieved simultaneously.