Precoder design for concatenating convolutional codes with generalized partial response channels

We study the performance of turbo codes for magnetic recording channels using realistic channel models. The main objective is to provide a benchmark for the future studies on the subject. We examine a variety of schemes including the use of parallel concatenated turbo codes with and without turbo equalization and convolutional codes with an interleaver (can be regarded as a serial concatenation of convolutional codes with the partial response channel). We investigate the effect of the (turbo) code block length, the use of different code structures, and the effects of precoding and the media noise on the system performance. We assume that the partial response target is EPR4.

Combination of turbo codes with partial response (PR) channels was comprehensively studied in magnetic recording systems. In this paper, we investigate a decoding scheme called 'iterative tri-stage decoding' for the turbo code as the outer code and the precoded PR channel as the inner code. This decoding scheme reduces the computational complexities by reducing the number of iterations without degrading the coding gain.

The partial response channel can be viewed as a rate-1 encoder in which the output alphabet differs from the input alphabet. In serially concatenated coding schemes, the partial response channel can serve as the inner encoder. Recent work has focused on convolutional codes and parity-check codes as the outer code in such serial concatenation schemes. This work focuses instead on the use of product codes as the outer code. Systems using both hard and soft decoding, with both iterative and non-iterative decoding, are presented. Simulation results are presented comparing performance of product codes which use Reed-Solomon and BCH codes as the constituent codes.

We study a serially interleaved concatenated code construction, where the outer code is a standard convolutional code, and the inner code is a recursive convolutional code of rate 1. We focus on the ubiquitous inner differential encoder (used, in particular, to resolve phase ambiguities), double differential encoder (used to resolve both phase and frequency ambiguities), and another rate 1 recursive convolutional code of memory 2. We substantiate analytically the rather surprising result, that the error probabilities corresponding to a maximum-likelihood (ML) coherently detected antipodal modulation over the additive white Gaussian noise (AWGN) channel for this construction are advantageous as compared to the stand-alone outer convolutional code. This is in spite of the fact that the inner code is of rate 1. The analysis is based on the tangential sphere upper bound of an ML decoder, incorporating the ensemble weight distribution (WD) of the concatenated code, where the ensemble is generated by all random and uniform interleavers. This surprising result is attributed to the WD thinning observed for the concatenated scheme which shapes the WD of the outer convolutional code to resemble more closely the binomial distribution (typical of a fully random code of the same length and rate). This gain is maintained regardless of a rather dramatic decrease, as demonstrated here, in the minimum distance of the concatenated scheme as compared to the minimum distance of the outer stand-alone convolutional code. The advantage of the examined serially interleaved concatenated code, given in terms of bit and/or block error probability which is decoded by a practical suboptimal decoder, over the optimally decoded standard convolutional code is demonstrated by simulations, and some insights into the performance of the iterative decoding algorithm are also discussed. Though we have investigated only specific constructions of constituent inner (rate 1) and outer codes, we trust, hinging on the rational of the arguments here, that these results extend to many other constituent convolutional outer codes and rate 1 inner recursive convolutional codes.

The partial response channel can be viewed as a rate-1 encoder in which the output alphabet differs from the input alphabet. In serially concatenated coding schemes, the partial response channel can serve as the inner encoder. Recent works on the application of turbo decoding techniques to partial response channels have used channel models to investigate system performance. While the models provided some insight into the performance of turbo decoding techniques, performance in a real system, with electronic and signal-dependent media noise, remained unknown. This paper will demonstrate the performance of turbo decoding techniques using real and modeled spinstand data. A simple serial concatenation system will be presented that uses as its outer code a single convolutional code and as its inner code the partial response channel. Significant gains will be shown over QMTR coded systems.

We present a systematic way to construct multilevel concatenated codes for partial response (PR) channels using: (1) a structured set partition (SSP) of multiple channel output sets and (2) a set of conventional block codes with different error correcting capabilities. A lower bound on the minimum squared Euclidean distance of the constructed codes is given. This bound is based on the interset minimal Euclidean distances of the SSP and the minimum Hamming distances of the used block codes. An example of SSP for the extended class 4 partial response channel (EPR4) is presented. Iterative suboptimal decoding, which combines Viterbi detection on the trellis of the PR channel with algebraic error detection/correction, can be applied to the constructed concatenated codes. Truncated versions of the iterative decoding scheme are simulated and compared with each other.

The dependence of MSE and marginalized BER on amplitude asymmetry for two partial response channels is reported. To eliminate the impact of variations in other head parameters, we varied the bias point of a given head using an external magnetic field. We find that the MSE degrades continuously as the magnitude of the amplitude asymmetry is increased; the marginalized BER is at most weakly dependant on amplitude asymmetry for magnitudes less than 15%. At larger magnitudes the channel performance is more severely impacted, Finally, we were able to simulate these results using a simplified model of the read channel.

This paper is concerned with the applications of majority-logic decodable nonbinary low-density parity-check (LDPC) codes to partial response (PR) channels. We propose a joint detection/decoding algorithm that works in an iterative manner by exchanging messages between the Viterbi detector and a generalized majority logic decoder (GMLGD). The Viterbi detector is implemented over a sectionalized trellis. The hard-decisions made by the Viterbi detector are then passed to the decoder. The decoder delivers as output estimates of each coded symbols, which will be utilized in next iteration to update the branch metrics of the trellis. Since the proposed algorithm requires only integer operations and finite field operations, it can be implemented with simple combinational logic circuits. Simulation results and complexity analysis show that, when compared with the conventional turbo equalizer implemented with the BCJR algorithm and the Q-ary sum-product algorithm (BCJR-QSPA), the proposed algorithm has a much lower complexity but suffers from a little performance degradation. So the proposed algorithm provides a good candidate for trade-offs between performance and complexity.

Transmitter and receiver filter design plays a key role in communication systems with partial response channels. For a partial response system, a nonmaximally decimated multirate filter bank can be used as a precoder. Based on the MMSE criterion, a closed-form solution to the joint transmitter-receiver optimization problem for noisy partial response channels is presented in this paper. It is found that the redundancy introduced by the nonmaximally decimated filter bank precoder can compensate for spectral nulls in partial response channels that can impair several known joint optimization methods. Simulation results corroborate the analysis that joint transmitter-receiver optimization can achieve significant performance improvement over inverse filter receiver and receiver optimization schemes.

Recording media written-in transition jitter noise can limit magnetic recording systems. Read channel intersymbol interference (ISI) modulates media jitter in a head/disk/channel, attenuating jitter at low linear density due to classical ISI but amplifying at high density. This paper assesses whether jitter modulation is significant in typical disk drive peak detect (PD) and partial response (PR) channel design, where maximum linear bit density is sought at a specified bit error rate (BER). Experimental isolated pulse data, transition noise, and electronic noise measurements from a 65 mm head/disk recording design are used as inputs to a Matlab simulation of PD and PR Lorentzian channel low pass and boost filters, PD differentiator filter, PR digital filter, and PD/PR bit detectors. The goal of such models is optimizing the designable PD and PR channel parameters for realistic maximum linear density prediction. However their complexity obscures the jitter modulation contribution, which is demonstrated here by the shift in the EER versus density curves, when the jitter modulation is mathematically turned off. Jitter modulation is thereby shown to be significant but not dominant in (1,7) PD density limits. About an order of magnitude BER loss due to jitter modulation can occur at typical PD linear channel bit densities P/sub 50//T/spl ap/2. Jitter amplification in PD channels can occur at channel density 3.8 (above usable PD limits). PR channels only amplify media jitter, degrading EER by about an order of magnitude at typical (0, k) PR4 densities, and degrading timing channel accuracy twice as fast as channel density increases. Lorentzian isolated pulse (ISOP) modeling is compared to recording physics mathematical pulses, and simulation statistical validity checks are made.

Towards the goal of achieving better error correction performance in data storage systems, iterative soft decoding of low density parity check (LDPC) codes and soft-decision decoding of Reed-Solomon (RS) codes have started receiving increasing research attention. However, even with increased computing power, complexities of soft-decision decoding algorithms are still too high for real products which require high throughput and small hardware area. Another problem is that the performance gains of those approaches are smaller for magnetic recording channels than they are for memoryless additive white Gaussian noise (AWGN) channels. We propose a new soft-decision decoding algorithm (based on the Chase algorithm), which takes advantage of pattern reliability instead of symbol reliability or bit reliability. We also present a modified Viterbi algorithm that provides probable error patterns with corresponding reliabilities. Simulation results of the proposed algorithms over the partial response (PR) channel show attractive performance gains. The proposed algorithm dramatically reduces the number of iterations compared to the conventional Chase2 algorithm over the PR channel.

This paper addresses the design of a protograph-based LDPC code which can approach the independent and uniformly distributed (i.u.d.) capacity of partial response channels. We propose a method to calculate the iterative decoding threshold of a joint graph between a protograph and the state structure of a partial response channel using the extrinsic information transfer (EXIT) chart. We then describe a simple method to search for a protograph code whose threshold is close to the i.u.d. capacity limit. This new class of codes is needed because experiments show a protograph that is capacity approaching in the AWGN channel may not perform well in partial response channels. In particular, protographs with punctured nodes are often used to produce good AWGN codes, but they perform poorly with the BCJR equalizer. Numerical results support our analysis.

We compare the asymptotic performance of parallel concatenated codes (PCCs) and serial concatenated codes (SCCs) over GF(4). We construct such codes using optimal nonbinary convolutional codes where optimality is in the sense of achieving the largest minimum distance for a fixed number of encoder states. Code rates of the form K/sub 0//(K/sub 0/ + 1) for K/sub 0/ = 1, 8, and 64 are considered. For all of these code rates, we find the minimum distance and its multiplicity for both concatenated code systems. This is accomplished by feeding the encoder with all possible weight-two and weight-three input information patterns and monitoring, at the output of the encoder, the weight of the corresponding codewords and their multiplicity. Our analytical results indicate that the SCC codes outperform their counterpart PCC codes with much lower complexity.

In this paper, a physical model of readback signals from a hybrid magneto-optical (MO) recording using magnetic flux detection is presented. The minimum mean-square error (MMSE) technique has been used to design generalized partial response (GPR) targets in order to facilitate the application of the trellis based channel detector. A PR target with integer coefficients is proposed as a good approximation to the GPR target. We also address the application of turbo codes and LDPC codes with iterative soft decoding structure to the hybrid magneto-optical (MO) recording channel. The effects of bloom, transition noise and giant magneto-resistive (GMR) sensor nonlinearity on the bit error rate (BER) performance are investigated both for uncoded partial response maximum likelihood (PRML) and iterative soft decoded recording channels. Our results for the hybrid recording channels indicate that turbo/LDPC codes of rate 16/17 can achieve coding gains of about 5 dB over PRML methods for a 10/sup -5/ target BER at the normalized density S=1.5. Turbo codes with iterative soft decoding are seen to be robust to transition noise.

Summary form only given. Following the work done by Ryan (Proc. ICC Int. Conf. on Communications, 1998), considerable research has been reported on the application of concatenated codes to partial response (PR) channels. Both parallel concatenated codes (PCC) and serial concatenated codes (SCC) have been studied. In all of those works, a linear model for the channel was assumed, which is a valid assumption for low recording densities. However, as the recording density increases, the channel becomes nonlinear where nonlinearity arises from closely spaced recorded transitions. In this paper, we combine concatenated coding and nonlinear equalization for PR channels that suffer from nonlinear distortion (or intersymbol interference) according to the setup shown in Fig 1. We extend the RAM-search/Viterbi-detector concept to the a posteriori probability (APP) and soft-output Viterbi algorithm (SOVA) detectors that are used with turbo-coded partial response channels. The modified APP and SOVA are called RS-APP and RS-SOVA, respectively.