Synchronization in digital communication systems: performance bounds and practical algorithms

Communication channels often transfer signals from different transmitters. To avoid interference the available frequency spectrum is divided into non-overlapping frequency bands (bandpass channels) and each transmitter is assigned to a different bandpass channel. The transmission of a signal over a bandpass channel requires a shift of its frequency-content to a frequency range that is compatible with the designated frequency band (modulation). At the receiver, the modulated signal is demodulated (frequency shifted back to the original frequency band) in order to recover the original signal. The modulation/demodulation process requires the presence of a locally generated sinusoidal signal at both the transmitter and the receiver. To enable a reliable information transfer, it is imperative that these two sinusoids are accurately synchronized. Recently, several powerful channel codes have been developed which enable reliable communication at a very low signal-to-noise ratio (SNR). A by-product of these developments is that synchronization must now be performed at a SNR that is lower than ever before. Of course, this imposes high requirements on the synchronizer design. This doctoral thesis investigates to what extent (performance bounds) and in what way (practical algorithms) the structure that the channel code enforces upon the transmitted signal can be exploited to improve the synchronization accuracy at low SNR