A Blind Sequential Monte Carlo Detector for OFDM Systems in the Presence of Phase Noise, Multipath Fading, and Channel Order Uncertainty

• Published in: IEEE transactions on signal processing 2007-09, Vol.55 (9), p.4581-4598
• Main Authors: Yee, D., Reilly, J.P., Kirubarajan, T.
• Format: Article
• Language: English
• Subjects: Algorithms; Applied sciences; Blind detection; Blinds; channel order; Channels; Detection, estimation, filtering, equalization, prediction; Detectors; Exact sciences and technology; Fading; frequency-selective fading; Information, signal and communications theory; Markov processes; Monte Carlo methods; Multiplexing; Noise; OFDM; Orthogonal Frequency Division Multiplexing; orthogonal-frequency-division multiplexing (OFDM); particle filtering; Phase detection; Phase frequency detector; Phase noise; sequential Monte Carlo (SMC); Signal and communications theory; Signal, noise; Sliding mode control; Studies; Telecommunications and information theory; Time varying systems; Uncertainty
• ISSN: 1053-587X; 1941-0476
• DOI: 10.1109/TSP.2007.896023

In this paper, an efficient detector is developed to address the blind detection problem for an orthogonal- frequency-division-multiplexing (OFDM) system in the presence of phase noise and unknown multipath fading, even with channel order that is possibly not known and time varying. The proposed maximum a posteriori detector is a combination of the sequential Monte Carlo (SMC) method and the variance reduction strategy known as Rao-Blackwellization. Being blind, the developed detector, does not rely on pilot tones for the detection of the transmitted data. However, as in most work found in the literature, the aforementioned detector, which we call the RB-SMC detector, invokes the assumption of a fixed and known channel order, which may be a limitation in a number of scenarios. Therefore, to relax this assumption, we model channel order uncertainty via a first-order Markov process and subsequently introduce appropriate extensions to the RB-SMC detector, thereby proposing a novel algorithm called the E-RB-SMC detector. The performance of the novel SMC-based detectors are validated through computer simulations. It is shown that the proposed SMC-based detectors achieve near bound performance. In terms of convergence speed, the proposed E-RB-SMC detector also shows the smallest acquisition time amongst the considered algorithms.