This project is an investigation of discrete time burst mode synchronization algorithms. The goal of the work is to minimize acquisition time in synchronization to enable applications to burst mode systems such as multi-user architectures. The primary design tool is the discrete time maximum likelihood estimator. Lower bounds are sought for symbol timing with an unknown phase or frequency and for the performance of joint symbol time and carrier phase and frequency estimation wit doppler. This will establish theoretical lower limits of the acquisition performance. Maximum likelihood estimator structures for discrete time burst mode synchronization systems will then be sought. The work will focus on recursive and nonrecursive discrete time implementations. Stochastic approximation techniques will be used in the solution of the maximum likelihood equations. This will result in digital phase lock loop systems with time varying gains, greatly improving the acquisition performance over constant gain realizations. Recursive structures will be developed for both joint and marginal synchronization of carrier phase and symbol timing. Statistical characterization of the acquisition process of the resulting estimators is a major emphasis. Markov theory and efficient Monte Carlo simulation are tools. Optimal techniques for detecting the start of a burst and optimal preamble designs will be developed and investigated.
