Low-cost code-aided ML timing recovery from turbo-coded QAM transmissions

Abstract

In this paper, we propose a new code-aided (CA) maximum likelihood (ML) approach for time synchronization in turbo-coded systems. The time delay estimate is refined at each turbo iteration owing to the increasingly accurate estimates for the log-likelihood ratios (LLRs) of the coded bits. The refined time delay estimate is then used by the matched filter (MF) in order to provide the soft-input soft-output (SISO) decoders with more reliable symbol-rate samples for the next turbo iteration. Simulation results show the remarkable performance improvements of CA estimation against the traditional non-data-aided (NDA) estimation scheme. Moreover, the new CA ML estimator (MLE) enjoys significant advantage in computational complexity over existing ML CA solutions.