Joint Channel Estimation and Multiuser Detection for SDMA/OFDM Based on Dual Repeated Weighted Boosting Search

• Published in: IEEE transactions on vehicular technology 2011-09, Vol.60 (7), p.3265-3275
• Main Authors: Jiankang Zhang, Sheng Chen, Xiaomin Mu, Hanzo, L.
• Format: Article
• Language: English
• Subjects: Applied sciences; Boosting; Channel estimation; Channels; Circuit properties; Complexity; Decoders; Detection, estimation, filtering, equalization, prediction; Detectors; Electric, optical and optoelectronic circuits; Electronic circuits; Electronics; Error correction; Error correction & detection; Estimators; Exact sciences and technology; Forward error correction; Information, signal and communications theory; Joint channel estimation and multiuser detection (JCEMUD); Joints; Mean square errors; Multiaccess communication; Multiplexing; Multiuser detection; OFDM; Orthogonal Frequency Division Multiplexing; orthogonal frequency-division multiplexing (OFDM); Reduction; repeated weighted boosting search (RWBS); Searching; Signal and communications theory; Signal convertors; Signal, noise; space-division multiple access (SDMA); Studies; Systems, networks and services of telecommunications; Telecommunications; Telecommunications and information theory; Transmission and modulation (techniques and equipments)
• ISSN: 0018-9545; 1939-9359
• DOI: 10.1109/TVT.2011.2161356

A joint channel estimation and multiuser detection (JCEMUD) scheme is proposed for multiuser multiple-input-multiple-output (MIMO) space-division multiple-access/orthogonal frequency-division-multiplexing (SDMA/OFDM) systems. We design a dual repeated weighted boosting search (DRWBS) scheme for JCEMUD, which is capable of providing "soft" outputs, which are directly fed to the forward error correction (FEC) decoder. The proposed DRWBS-JCEMUD scheme iteratively estimates the channel impulse responses and detects the users' transmitted signals while exploiting the error correction capability of an FEC decoder to iteratively exchange information between the detector and the estimator. Furthermore, the proposed DRWBS-JCEMUD scheme is capable of providing the log-likelihood ratios of the coded bits at low computational complexity (comparable with the single-user scenario), which can directly be fed to the FEC decoder. The simulation results demonstrate that the proposed DRWBS-JCEMUD scheme is capable of attaining a mean square error performance close to that of the ideal scenario of the least-square channel estimator associated with 100% pilot overhead and narrows the discrepancy with respect to the optimal maximum-likelihood (ML) MUD associated with perfect channel knowledge. As an example, at E b /N 0 = 10 dB, a factor-of-0.756 complexity reduction was achieved at the cost of a 1-dB performance penalty, in comparison with the ML-MUD.