Phase domain maximum likelihood carrier recovery: framework and application in wireless TDMA systems

Abstract

In this paper, the maximum likelihood (ML) principle in the phase domain is employed to provide a solution to the carrier recovery problem appropriate for reverse channel point-to-multipoint TDMA wireless communication systems. Frequency and phase estimators are derived exhibiting the minimum variance unbiased estimation (MVUE) property under an assumption on the received signal power. Extensions of the basic estimation algorithm are provided for the cases of pattern-based carrier recovery, decision-directed as well as non-decision-directed carrier recovery based on random data. These three techniques are then used in a burst data transmission platform employing either QPSK or 16-QAM modulation format. A scheme for simultaneous clock and carrier recovery is also considered. Root raised cosine transmit and receive filtering is used. Extensive simulations have shown that a 20-symbol preamble is sufficient for providing reliable system parameter recovery in all tested cases. Specifically for the 16-QAM modulated signal, achieving joint carrier and clock recovery with a preamble as short as 20 symbols is an important result, very beneficial for bandwidth-constrained TDMA system design.