Iterative Code-Aided ML Phase Estimation and Phase Ambiguity Resolution
- PDF / 773,894 Bytes
- 8 Pages / 600 x 792 pts Page_size
- 10 Downloads / 209 Views
Iterative Code-Aided ML Phase Estimation and Phase Ambiguity Resolution Henk Wymeersch Digital Communications Research Group, Department of Telecommunications and Information Processing, Ghent University, Sint-Pietersnieuwstraat 41, 9000 Ghent, Belgium Email: [email protected]
Marc Moeneclaey Digital Communications Research Group, Department of Telecommunications and Information Processing, Ghent University, Sint-Pietersnieuwstraat 41, 9000 Ghent, Belgium Email: [email protected] Received 29 September 2003; Revised 25 May 2004 As many coded systems operate at very low signal-to-noise ratios, synchronization becomes a very difficult task. In many cases, conventional algorithms will either require long training sequences or result in large BER degradations. By exploiting code properties, these problems can be avoided. In this contribution, we present several iterative maximum-likelihood (ML) algorithms for joint carrier phase estimation and ambiguity resolution. These algorithms operate on coded signals by accepting soft information from the MAP decoder. Issues of convergence and initialization are addressed in detail. Simulation results are presented for turbo codes, and are compared to performance results of conventional algorithms. Performance comparisons are carried out in terms of BER performance and mean square estimation error (MSEE). We show that the proposed algorithm reduces the MSEE and, more importantly, the BER degradation. Additionally, phase ambiguity resolution can be performed without resorting to a pilot sequence, thus improving the spectral efficiency. Keywords and phrases: turbo synchronization, phase estimation, phase ambiguity resolution, EM algorithm.
1.
INTRODUCTION
In packet-based communications, frames arrive at the receiver with an unknown carrier phase. When phase estimation (PE) is performed by means of a conventional non-dataaided (NDA) algorithm [1], the resulting estimate exhibits a phase ambiguity, due to the rotational symmetries of the signalling constellation. Phase ambiguity resolution (PAR) can be accomplished by a data-aided (DA) algorithm that exploits the presence of a known pilot sequence in the transmitted data stream [2]. The need for PAR can be removed by using differential encoding, which however results in a BER degradation, and requires significant changes to the decoder in case of iterative demodulation/decoding [3]. Since a phase ambiguity resolution failure gives rise to the loss of an entire packet, its probability of occurrence should be made sufficiently small. At the same time, the pilot sequence must not be too long as it reduces the spectral efficiency of the system. Although conventional estimation algorithms perform well for uncoded systems, a different approach needs to be taken when powerful error-correcting codes are used. These codes operate typically at low SNR, making the estimation
process more difficult. By exploiting the knowledge of certain code properties, a more accurate estimate may be obtained. In [4], by approximating the log-likelihood function,
Data Loading...
