This paper presents a multiplierless discrete cosine transforms (DCT) design with approximate canonical signed digit (CSD) encoding for image processing. Two approximation strategies on CSD encoding are proposed in constant multiplication. Based on these two coding approaches, an approximate DCT architecture is presented by taking advantage of the correlation between adjacent pixels of image data. Higher frequency coefficients are gradually ignored from the calculation due to the energy compaction property. Four approximate DCT architectures are thus proposed representing different accuracy levels. The proposed DCTs are implemented using 0.18μm standard CMOS process. The simulation results indicate that the proposed ADCT reduces 51.5% power and 30.0% area with a PSNR penalty of 1.6dB when compared with the traditional design. For lossy applications which allow lower computational accuracy, ADCT-III achieves 70.0% and 43.9% reduction on the power consumption and area, respectively, at a cost of 11.7 dB accuracy loss.

A code-matched interleaver design for turbo codes in which a particular interleaver is constructed to match the code weight distribution is proposed. The design method is based on the code distance spectrum. The low weight paths in the code trellis which give large contributions to the error probability in the signal-to-noise ratio region of interest for practical communication systems are eliminated so that they do not appear in the overall code trellis after interleaving. The proposed interleaver improves the code error performance at moderate to high signal-to-noise ratio and considerably increases the asymptotic slope of the error probability curves.

The performance of parallel and serial concatenated codes on frequency-nonselective fading channels is considered. The analytical average upper bounds of the code performance over Rician channels with independent fading are derived. Furthermore, the log-likelihood ratios and extrinsic information for maximum a posteriori (MAP) probability and soft-output Viterbi algorithm (SOVA) decoding methods on fading channels are developed. The derived upper bounds are evaluated and compared to the simulated bit-error rates over independent fading channels. The performance of parallel and serial codes with MAP and SOVA iterative decoding methods, with and without channel state information, is evaluated by simulation over independent and correlated fading channels. It is shown that, on correlated fading channels, the serial concatenated codes perform better than parallel concatenated codes. Furthermore, it has been demonstrated that the SOVA decoder has almost the same performance as the MAP decoder if ideal channel state information is used on correlated Rayleigh fading channels.

A new algorithm for optimal design of joint source-channel coding over coded communication systems is presented. The source coder and decoder (codec) based on the vector quantization (VQ) technique are optimized for the coded system by minimizing the end-to-end system distortion. The channel codec is realized with a turbo code. The proposed algorithm is derived based on a discrete-input analog-output additive white Gaussian noise (AWGN) channel model. Maximum a posteriori (MAP) are used for soft-output channel decoding. The soft output signals are applied directly to the designed optimal source decoder for source decoding. Applications of the proposed algorithm for wireless low-bit-rate speech transmission is also given. Simulation results show that, the proposed algorithm can objectively gain 4-5 dB improvement of decoded source SNR (DSSNR) for a first-order Gauss-Markov source, and subjectively achieve 2-3 mean-opinion scores (MOS) enhancement for the decoded speech quality relative to the traditional system with hard decoding outputs, especially over the low SNR channels.

The impact of the distance spectrum and interleaver structure on the bit error probability of turbo codes is considered. A new turbo code design method for Gaussian channels is presented. The proposed method combines a search for good component codes with interleaver design. The optimal distance spectrum is used as the design criterion to construct good turbo component codes at low signal-to-noise ratios (SNRs). In addition, an interleaver design method is proposed. This design improves the code performance at high SNR. Search for good component codes at low SNR is combined with a code matched interleaver design. This results in new turbo codes with a superior error performance relative to the best known codes at both low and high SNR. The performance is verified by both analysis and simulation.