ESTIMATION AND IMPROVEMENT OF THE PERFORMANCE OF COMMUNICATION SYSTEMS BASED ON LDPC CODES

DoctorThe objective of this thesis is to estimate and improve the performance ofcommunication systems employing low-density parity-check (LDPC) codes. Wefirstly propose a hardware-efficient adaptive modulation and coding (AMC) schemewhich uses only a single LDPC encoder/decoder with irregular modulation. TwoGray-mapped quadrature-amplitude modulations are employed within a code-word so that the transmission rates of the proposed AMC can be controlledeven though the rate of the employed LDPC code is fixed. The rate and degreedistribution of the employed LDPC code are selected properly for the proposedscheme. We also optimize the bit-to-symbol mapping for the LDPC code over ir-regular modulation at each transmission rate. The performance of the proposedscheme is compared with that of the AMC scheme with a dedicated LDPC codeat each transmission rate.Our second consideration is to estimate the performance of LDPC-coded sys-tems. We propose a novel importance sampling (IS) scheme to estimate the lowerror-rate performance of LDPC codes over Rayleigh fading channels. The pro-posed scheme exploits the structural weakness of LDPC codes due to trappingsets (TSs). The Rayleigh fading distribution on the bits belonging to a TS is bi-ased by parameter scaling (PS), while the noise distribution on them is biasedvia mean translation (MT) according to their fading coefficients. The biases inPS and MT are determined so that the variance of the proposed IS estimator is minimized. The proposed IS scheme is compared with the Monte Carlo (MC)simulator and other IS schemes modified from the conventional IS scheme em-ployed for performance estimation of LDPC codes over an AWGN channel.In the last part of this thesis, we propose a novel LDPC-coded frequency-domain equalizer (FDE) for single-carrier block transmission over a doubly-selective channel. The proposed LDPC-coded FDE partitions a received blockinto multiple subblocks in order to reduce the effect of the time variation ofthe channel. The nonuniformity in the error statistics of a subblock-based FDEis analyzed, when a one-tap minimum mean-squared error (MMSE) equalizeris employed. The nonuniformity is exploited to remove the drawback of sub-block processing so that the advantage of a subblock-based FDE is maximized.Numerical results show that the proposed FDE has much better performancethan other known subblock-based FDEs and has a robust performance overhigh Doppler environments