Iterative Synchronization at Low Signal-to-Noise Ratios for Wireless Communications

Within this thesis, transmission over a wireless communication channel in low signal-to-noise ratio (SNR) environments is considered. Such environments demand for a strong error-correcting code in order to ensure reliable transmission. Strong error-correcting codes, such as Turbo codes or LDPC codes, are especially sensitive to synchronization errors. Additionally, accurate synchronization is achallenging task in low SNR environments. In such a scenario, synchronization is conventionally pursued with data-aided algorithms that make use of pilot symbols. A drawback of these algorithms is that their estimation accuracy scales with the number of pilot symbols. Hence, highly accurate synchronization at low SNR comes at the cost of a large overhead of pilot symbols, and, thus, a lower bandwidth and power efficiency. An option to overcome this dilemma is code-aided synchronization. Code-aided algorithms exploit the error-correcting code for synchronization purpose. They are implemented by establishing a feedback loop from the channel decoder to the estimation unit, such that both units can iteratively exchange information, and yield iteratively improving results.

Conventional data-aided synchronization algorithms and code-aided synchronization algorithms are considered jointly in this thesis. In particular, the question of how accurate initial synchronization needs to be in order to ensure (reasonably fast) convergence of the codeaided synchronization unit is investigated. Furthermore, performance limits of code-aided synchronization are pointed out.