Markov model aided decoding for image transmission usingsoft-decision-feedback

Soft-decision-feedback MAP decoders are developed for joint source/channel decoding (JSCD) which uses the residual redundancy in two-dimensional sources. The source redundancy is described by a second order Markov model which is made available to the receiver for row-by-row decoding, wherein the output for one row is used to aid the decoding of the next row. Performance can be improved by generalizing so as to increase the vertical depth of the decoder. This is called sheet decoding, and entails generalizing trellis decoding of one-dimensional data to trellis decoding of two-dimensional data (2-D). The proposed soft-decision-feedback sheet decoder is based on the Bahl algorithm, and it is compared to a hard-decision-feedback sheet decoder which is based on the Viterbi algorithm. The method is applied to 3-bit DPCM picture transmission over a binary symmetric channel, and it is found that the soft-decision-feedback decoder with vertical depth V performs approximately as well as the hard-decision-feedback decoder with vertical depth V+1. Because the computational requirement of the decoders depends exponentially on the vertical depth, the soft-decision-feedbark decoder offers significant reduction in complexity. For standard monochrome Lena, at a channel bit error rate of 0.05, the V=1 and V=2 soft-decision-feedback decoder JSCD gains in RSNR are 5.0 and 6.3 dB, respectively.     

Smoothed DPCM codes

Interpolative differential pulse code modulation (IDPCM) is known to outperform DPCM at rate 1 b/sample for several synthetic source models. The authors use minimum-mean-squared error (MMSE) fixed-lag smoothing with DPCM to develop a code generator using delayed decoding. This smoothed DPCM (SDPCM) code generator is compared to DPCM and IDPCM code generators at rates 1 and 2 b/sample for tree coding several synthetic sources and to a DPCM code generator at 2 b/sample for speech sources     

Combining hidden Markov source models and parallel concatenatedcodes

We present here a framework for modifying a decoder for parallel concatenated codes to incorporate a general hidden Markov source model. This allows the receiver to utilize the statistical characteristics of the source during the decoding process, and leads to significantly improved performance relative to systems in which source statistics are not exploited. One of the constituent decoders makes use of a modified trellis which jointly describes the source and the encoder. The number of states in this modified trellis is the product of the number of states in the hidden Markov source and the number of states in the encoder     

A 500-Mb/s soft-output Viterbi decoder

Two eight-state 7-bit soft-output Viterbi decoders matched to an EPR4 channel and a rate-8/9 convolutional code are implemented in a 0.18-/spl mu/m CMOS technology. The throughput of the decoders is increased through architectural transformation of the add-compare-select recursion, with a small area overhead. The survivor-path decoding logic of a conventional Viterbi decoder register exchange is adapted to detect the two most likely paths. The 4-mm/sup 2/ chip has been verified to decode at 500 Mb/s with 1.8-V supply. These decoders can be used as constituent decoders for Turbo codes in high-performance applications requiring information rates that are very close to the Shannon limit.     

Joint turbo decoding and estimation of hidden Markov sources

We describe a joint source-channel scheme for modifying a turbo decoder in order to exploit the statistical characteristics of hidden Markov sources. The basic idea is to treat the trellis describing the hidden Markov source as another constituent decoder which exchanges information with the other constituent decoder blocks. The source block uses as extrinsic information the probability of the input bits that is provided by the constituent decoder blocks. On the other hand, it produces a new estimation of such a probability which will be used as extrinsic information by the constituent turbo decoders. The proposed joint source-channel decoding technique leads to significantly improved performance relative to systems in which source statistics are not exploited and avoids the need to perform any explicit source coding prior to transmission. Lack of a priori knowledge of the source parameters does not degrade the performance of the system, since these parameters can be jointly estimated with turbo decoding     

Graph-based iterative decoding algorithms for parity-concatenatedtrellis codes

We construct parity-concatenated trellis codes in which a trellis code is used as the inner code and a simple parity-check code is used as the outer code. From the Tanner-Wiberg-Loeliger (1981, 1996) graph representation, several iterative decoding algorithms can be derived. However, since the graph of the parity-concatenated code contains many short cycles, the conventional min-sum and sum-product algorithms cannot achieve near-optimal decoding. After some simple modifications, we obtain near-optimal iterative decoders. The modifications include either (a) introducing a normalization operation in the min-sum and sum-product algorithms or (b) cutting the short cycles which arise in the iterative Viterbi algorithm (IVA). After modification, all three algorithms can achieve near-optimal performance, but the IVA has the least average complexity. We also show that asymptotically maximum-likelihood (ML) decoding and a posteriori probability (APP) decoding can be achieved using iterative decoders with only two iterations. Unfortunately, this asymptotic behavior is only exhibited when the bit-energy-to-noise ratio is above the cutoff rate. Simulation results show that with trellis shaping, iterative decoding can perform within 1.2 dB of the Shannon limit at a bit error rate (BER) of 4×10^-5 for a block size of 20000 symbols. For a block size of 200 symbols, iterative decoding can perform within 2.1 dB of the Shannon limit     

Parallel Viterbi decoding methods for uncontrollable andcontrollable sources

The parallel Viterbi decoding is discussed for two different cases: uncontrollable sources and controllable sources. For general, uncontrollable Markov processes, a previously known parallel method is extended to a hierarchical parallel decoding approach, which achieves a lower latency. For controllable Markov sources in telecommunications applications, new parallel decoding methods are obtained by controlling the source processes in appropriate ways. The focus is on the parallel decoding methods for controllable sources because these methods have zero processing overhead. Because the methods modify the coding process, they bring positive changes to framing and negative changes to latency and code performance. However, one can adjust the parameters of the methods to make the degradation negligible. Because of their low overhead, the methods are most attractive for high-speed decoders for convolutional and trellis codes, and they are also applicable to other sequential algorithms for suboptimal decoding and estimation of complex Markov sources     

Joint Source and Channel Coding using Punctured Ring Convolutional Coded CPM

In this paper, a novel trellis source encoding scheme based on punctured ring convolutional codes is presented. Joint source and channel coding (JSCC) using trellis coded continuous phase modulation (CPM) with punctured convolutional codes over rings is investigated. The channels considered are the additive white gaussian noise (AWGN) channel and the Rayleigh fading channel. Optimal soft decoding for the proposed JSCC scheme is studied. The soft decoder is based on the a posteriori probability (APP) algorithm for trellis coded CPM with punctured ring convolutional codes. It is shown that these systems with soft decoding outperform the same systems with hard decoding especially when the systems operate at low to medium signal-to-noise ratio (SNR). Furthermore, adaptive JSCC approaches based on the proposed source coding scheme are investigated. Compared with JSCC schemes with fixed source coding rates, the proposed adaptive approaches can achieve much better performance in the high SNR region. The novelties of this work are the development of a trellis source encoding method based on punctured ring convolutional codes, the use of a soft decoder, the APP algorithm for the combined systems and the adaptive approaches to the JSCC problem.     

Low-complexity iterative joint source-channel decoding for variable-length encoded Markov sources

In this paper, we present a novel packetized bit-level decoding algorithm for variable-length encoded Markov sources, which calculates reliability information for the decoded bits in the form of a posteriori probabilities (APPs). An interesting feature of the proposed approach is that symbol-based source statistics in the form of the transition probabilities of the Markov source are exploited as a priori information on a bit-level trellis. This method is especially well-suited for long input blocks, since in contrast to other symbol-based APP decoding approaches, the number of trellis states does not depend on the packet length. When additionally the variable-length encoded source data is protected by channel codes, an iterative source-channel decoding scheme can be obtained in the same way as for serially concatenated codes. Furthermore, based on an analysis of the iterative decoder via extrinsic information transfer charts, it can be shown that by using reversible variable-length codes with a free distance of two, in combination with rate-1 channel codes and residual source redundancy, a reliable transmission is possible even for highly corrupted channels. This justifies a new source-channel encoding technique where explicit redundancy for error protection is only added in the source encoder.     

Iterative maximum-likelihood trellis decoding for block codes

An iterative trellis search technique is described for the maximum-likelihood (ML) soft decision decoding of block codes. The proposed technique derives its motivation from the fact that a given block code may be a subcode for a parent code whose associated trellis has substantially fewer edges. Through the use of list-Viterbi (1967) decoding and an iterative algorithm, the proposed technique allows for the use of a trellis for the parent code in the ML decoding of the desired subcode. Complexity and performance analyses, as well as details of potential implementations, indicate a substantial reduction in decoding complexity for linear block codes of practical length while achieving ML or near-ML soft decision performance     

DPCM picture transmission over noisy channels with the aid of aMarkov model

The problem of DPCM picture transmission over noisy channels is considered. It is well known that DPCM systems are very sensitive to channel errors. The goal in this work is to build robustness against channel errors. Three methods are proposed in this paper and are obtained by modeling the encoded signal as a Markov sequence. First, an optimum method for decoding correlated sequences is derived, and it is shown to require Viterbi decoding. Then, a modified MAP method (MMAP) for Markov sequences is described. A maximal signal-to-noise (MSNR) receiver for DPCM systems is also developed that minimizes the distortion power due to channel errors. The appropriate cost matrix for this receiver is computed. These methods are applied to DPCM picture transmission over noisy channels and are compared with a another method. The SNR graphs, as well as subjective examination of the received pictures, demonstrate that the proposed procedures are quite effective and superior to that method. Among the proposed methods, the MSNR receiver was found to be more effective than the others for a given order of the Markov model. It is observed that the proposed methods are most beneficial for low detail pictures.     

SNR mismatch and online estimation in turbo decoding

Iterative decoding of turbo codes, as well as other concatenated coding schemes of similar nature, requires knowledge of the signal-to-noise ratio (SNR) of the channel so that proper blending of the a posteriori information of the separate decoders is achieved. We study the sensitivity of decoder performance to misestimation of the SNR, and propose a simple online scheme that estimates the unknown SNR from each code block, prior to decoding. We show that this scheme is sufficiently adequate in accuracy to not appreciably degrade the performance     

A Reconfigurable TDMP Decoder for Raptor Codes

A Raptor code is a concatenation of a fixed rate precode and a Luby-Transform (LT) code that can be used as a rateless error-correcting code over communication channels. By definition, Raptor codes are characterized by irregularity features such as dynamic rate, check-degree variability, and joint coding, which make the design of hardware-efficient decoders a challenging task. In this paper, serial turbo decoding of architecture-aware Raptor codes is mapped into sequential row processing of a regular matrix by using a combination of code enhancements and architectural optimizations. The proposed mapping approach is based on three basic steps: (1) applying systematic permutations on the source matrix of the Raptor code, (2) confining LT random encoding to pseudo-random permutation of messages and periodic selection of row-splitting scenarios, and (3) developing a reconfigurable parallel check-node processor that attains a constant throughput while processing LT- and LDPC-nodes of varying degrees and count. The decoder scheduling is, thus, made simple and uniform across both LDPC and LT decoding. A serial decoder implementing the proposed approach was synthesized in 65 nm, 1.2 V CMOS technology. Hardware simulations show that the decoder, decoding a rate-0.4 code instance, achieves a throughput of 36 Mb/s at SNR of 1.5 dB, dissipates an average power of 27 mW and occupies an area of 0.55 mm².     

Combined MMSE interference suppression and turbo coding for acoherent DS-CDMA system

The performance of a turbo-coded code division multiaccess system with a minimum mean-square error (MMSE) receiver for interference suppression is analyzed on a Rayleigh fading channel. In order to accurately estimate the performance of the turbo coding, two improvements are proposed on the conventional union bounds: the information of the minimum distance of a particular turbo interleaver is used to modify the average weight spectra, and the tangential bound is extended to the Rayleigh fading channel. Theoretical results are derived based on the optimum tap weights of the MMSE receiver and maximum-likelihood decoding. Simulation results incorporating iterative decoding, RLS adaptation, and the effects of finite interleaving are also presented. The results show that in the majority of the scenarios that we are concerned with, the MMSE receiver with a rate-1/2 turbo code will outperform a rate-1/4 turbo code. They also show that, for a bit error rate lower than 10^-3, the capacity of the system is increased by using turbo codes over convolutional codes, even with small block sizes     

A constraint-length based modified Viterbi algorithm with adaptiveeffort

A modified Viterbi (1971) algorithm for convolutional codes is described that provides for signal-to-noise ratio (SNR) adaptive computational effort. The algorithm has three levels of prioritized effort. Movement from one level to the next is controlled by parameters that can be selected according to desired output bit error rate performance. For 3-bit soft decision detected signals, a coding gain within 0.06 dB of Viterbi at a 3-dB SNR is achieved for the same constraint-length code with modest parameter values and computational effort. At values of SNR above 6 dB, the algorithm decodes with very low computational effort. Effort levels are controlled by spanning the decoding trellis in steps that are one constraint-length long     

Noncoherent demodulation techniques for trellis coded M-DPSKsignals

Considers noncoherent decoding of trellis coded M-DPSK transmission, and attempts to unify and contrast the approaches to this problem which have been previously reported. In the second part of the paper, the maximum likelihood block decoders for the AWGN channel are outlined, first assuming that the decoder has no knowledge of the state of the encoder at the beginning of the block, and then assuming that it knows the state. Because these exhaustive decoders suffer from an exponential growth in complexity with block length, two simplified decoders are proposed that approach the performance of the exhaustive decoders with a much more manageable complexity. Through a union bound analysis and computer simulations, it is shown that it is possible to achieve large gains over uncoded systems with a very reasonable decoder complexity     

Parallel high-throughput limited search trellis decoder VLSI design

Limited search trellis decoding algorithms have great potentials of realizing low power due to their largely reduced computational complexity compared with the widely used Viterbi algorithm. However, because of the lack of operational parallelism and regularity in their original formulations, the limited search decoding algorithms have been traditionally ruled out for applications demanding very high throughput. We believe that, through appropriate algorithm and hardware architecture co-design, certain limited search trellis decoding algorithms can become serious competitors to the Viterbi algorithm for high-throughout applications. Focusing on the well-known T-algorithm, this paper presents techniques at the algorithm and VLSI architecture levels to design fully parallel T-algorithm limited search trellis decoders. We first develop a modified T-algorithm, called SPEC-T, to improve the algorithmic parallelism. Then, based on the conventional state-parallel register exchange Viterbi decoder, we develop a parallel SPEC-T decoder architecture that can effectively transform the reduced computational complexity at the algorithm level to the reduced switching activities in the hardware. We demonstrate the effectiveness of the SPEC-T design solution in the context of convolutional code decoding. Compared with state-parallel register exchange Viterbi decoders, the SPEC-T convolutional code decoders can achieve almost the same throughput and decoding performance, while realizing up to 56% power savings. For the first time, this work provides an approach to exploit the low power potential of the T-algorithm in very high throughput applications.     

Design and decoding of optimal high-rate convolutional codes

This correspondence deals with the design and decoding of high-rate convolutional codes. After proving that every (n,n-1) convolutional code can be reduced to a structure that concatenates a block encoder associated to the parallel edges with a convolutional encoder defining the trellis section, the results of an exhaustive search for the optimal (n,n-1) convolutional codes is presented through various tables of best high-rate codes. The search is also extended to find the "best" recursive systematic convolutional encoders to be used as component encoders of parallel concatenated "turbo" codes. A decoding algorithm working on the dual code is introduced (in both multiplicative and additive form), by showing that changing in a proper way the representation of the soft information passed between constituent decoders in the iterative decoding process, the soft-input soft-output (SISO) modules of the decoder based on the dual code become equal to those used for the original code. A new technique to terminate the code trellis that significantly reduces the rate loss induced by the addition of terminating bits is described. Finally, an inverse puncturing technique applied to the highest rate "mother" code to yield a sequence of almost optimal codes with decreasing rates is proposed. Simulation results applied to the case of parallel concatenated codes show the significant advantages of the newly found codes in terms of performance and decoding complexity.     

Universal trellis codes for diagonally layered space-time systems

Foschini's diagonally layered space-time transmission system known as D-BLAST is an architecture designed for a Rayleigh fading environment using multiple element antenna arrays at both the transmit and receive sites to achieve very high spectral efficiencies. In this paper, we propose a simple coding technique for D-BLAST that uses a single trellis code with finite-traceback Viterbi decoding. We examine the performance of universal trellis codes that are designed to have a distance structure that is matched to the periodic signal-to-noise ratio (SNR) variation of the channel created by D-BLAST, under the assumption that the channel is static during one burst but may change from burst to burst. We show that a universal 64-state trellis code on a 2/spl times/2 D-BLAST system with long enough block lengths displays universal behavior working on almost every 2/spl times/2 channel with at least the mutual information required by a standard 64-state AWGN trellis code. The only 2/spl times/2 channel where more mutual information is required is a certain rotation of the zero eigenvalue channel. We also present 4/spl times/4 and 8/spl times/8 examples.     

Efficient bounded-distance decoding of the hexacode and associateddecoders for the Leech lattice and the Golay code

Two soft-decision decoding algorithms for the (6, 3, 4) quaternary code hexacode are presented. Both algorithms realize half the minimum Euclidean distance of the code. The proposed algorithms are most practical. In using them, bounded-distance decoding of the Golay code and the Leech lattice are performed with at most 187 and 519 real-number operations respectively. Compare this to 651, respectively 3595, operations required by the best known maximum likelihood decoders (Vardy and Be'ery, 1991, 1993), and 431, respectively 1007, operations required by the bounded-distance decoders (Amrani et al., 1994). We present some simulation results for the proposed Leech lattice decoders revealing near-optimal performance. A comparison to known trellis codes is also provided