A Low-Complexity Iterative Phase Noise Tracker for Bit-Interleaved Coded CPM Signals in AWGN

• Published in: IEEE transactions on signal processing Vol. 59; no. 9; pp. 4271 - 4285
• Main Authors: Noels, N., Moeneclaey, M., Simoens, F., Delaruelle, D.
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01.09.2011; Institute of Electrical and Electronics Engineers; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Applied sciences; Belief propagation; bit-interleaved coded modulation; Coding, codes; Coherence; continuous-phase modulation (CPM); Decisions; Detection, estimation, filtering, equalization, prediction; Detectors; Estimation; Exact sciences and technology; factor graph; Information, signal and communications theory; iterative estimation and detection; Iterative methods; Modulation; Modulation, demodulation; Modules; Noise; Phase locked loops; Phase noise; phase noise tracking; Receivers; Signal and communications theory; Signal, noise; Studies; sum-product algorithm; Symbols; Telecommunications and information theory; Tin; turbo carrier synchronization; Phase noise; factor graph; Parameter estimation; sum-product algorithm; AWGN; Error estimation; continuous-phase modulation (CPM); Iterative method; phase noise tracking; Implementation; Degradation; Credal approach; A posteriori estimation; Soft decision; iterative estimation and detection; turbo carrier synchronization; Synchronization; Computational complexity; Phase locked loop; Phase estimation; Signal processing; Interleaved codes; Maximum likelihood; bit-interleaved coded modulation; EM algorithm; Continuous phase modulation; Belief propagation
• ISSN: 1053-587X; 1941-0476
• DOI: 10.1109/TSP.2011.2159498

This paper considers iterative detection of bit-interleaved coded continuous phase modulation in the presence of both phase noise (PN) and additive white Gaussian noise (AWGN). The proposed receiver iterates between a detection module and an estimation module. The detection module operates according to the sum-product algorithm and the factor graph framework in order to perform coherent maximum a posteriori bit detection in AWGN, using a PN estimate provided by the estimation module. The latter module, which results from the expectation-maximization algorithm for maximum likelihood estimation of the unknown PN samples, is implemented as a smoothing phase-locked loop that uses soft decisions provided by the detector. The separation between the detection and the estimation modules allows the use of an off-the-shelf (coherent) bit detector. The technique is further characterized by a very low computational complexity, a small error performance degradation and a small number of overhead symbols.