Sayısal haberleşme sistemlerinde gürbüz kestirici tabanlı alıcı tasarımı

In this thesis, baseband and band-pass waveform design is proposed to transmit and receive binary information under non-Gaussian impulsive noise environment. The additive channel noise is modelled by alpha-stable distribution. Differing from the previous studies, the signal detection problem under alpha-stable noise is analysed by considering both symmetrical and skewed cases. It is shown that signal detection performance decreases when the noise exhibits asymmetric behaviour. Once the destructive effect of skewness of the noise on communication is determined, two strategies are proposed to reduce the skewness of the channel noise. The first attempt is to add alpha-stable noise having the same impulsiveness and opposite skewness at the input of the receiver in order to symmetrize the resultant noise. This operation corresponds to stochastic resonance phenomena. As the second approach, the robust estimators are utilized to increase the signal detection performance under skewed impulsive noise by introducing novel waveform design which is primarily based on converting signal to be transmitted into an intentionally antipodal waveform. The receiver undoes the same operation and uses robust estimators to reduce the skewed alpha-stable noise on the deterministic signal to recover the transmitted binary information even if the channel noise distribution is not known in advance. In order to expose the improvement of the proposed methods, coherent detection and non-coherent differential modulation based waveform design are proposed to illustrate the enhancement in bit error rate, respectively. The antipodal waveform design provides digital communication for not only skewed alpha-stable noise but also any non-Gaussian noise having symmetric and/or skewed distribution.