Rathnakumar Radhakrishnan

Channel Models and Detectors for Two-Dimensional Magnetic Recording

Two-dimensional magnetic recording (TDMR) is a novel recording architecture intended to support densities beyond those of conventional recording systems. The gains from TDMR come primarily from more powerful coding and signal processing algorithms that allow the bits to be packed more tightly on the disk, and yet be retrieved with acceptable error rates. In this paper, we present some preliminary results for an advanced channel model based on micromagnetic simulations, coined the Grain Flipping Probability model. This model requires a one-time computationally complex characterization phase, but subsequently provides fast and accurate two-dimensional (2-D) readback waveforms that include effects captured from micromagnetic simulations and the statistical effects derived from the granularity of the recording medium. We also show the performance of several detectors over a pre-existing TDMR channel model directly as a function of channel density rather than the signal-to-noise ratio (SNR).

2-D Magnetic Recording: Read Channel Modeling and Detection

Two-dimensional magnetic recording (TDMR) is a novel storage architecture that, in theory, can achieve a density of up to 10 Tb/in2. It uniquely differs from other proposed next-generation architectures because of its reliance on sophisticated 2-D signal-processing algorithms. Recently, a number of contributions have been made in the development of read-channel models and detectors for TDMR systems. In this paper, we provide a detailed review on all important read-channel models under consideration. Our discussion focuses on the purpose of each model, placing a special emphasis on the suitability of the Voronoi model for the purpose of designing detectors. We also propose several detection schemes for TDMR based on the Voronoi model and present some numerical results.

Read Channel Modeling for Detection in Two-Dimensional Magnetic Recording Systems

In this paper, we describe a read channel model for detector design for two-dimensional magnetic recording (TDMR) system, a novel strategy for recording at upto 10 Tb/in2. We describe a scheme for (1) modeling of the recording medium, (2) modeling of the write/readback process, and (3) an experimental method for the characterization of noise in the TDMR channel, occurring due to irregularities in the bit-boundaries in the recording medium, that can be used for detection purposes.

LDPC Decoding Strategies for Two-Dimensional Magnetic Recording

In this paper, we propose a linear programming (LP) decoding scheme for binary error-erasure channel for use in two-dimensional magnetic recording. We compare the performance of this decoding scheme with other decoding schemes like LP decoding for BSC and belief-propagation decoding. Also, we compare the effect of variance of grain-area in the medium on the bit-error rates of various decoding schemes.

Transition Response Characteristics of Heat-Assisted Magnetic Recording and Their Performance With MTR Codes

The response of a heat-assisted magnetic recording (HAMR) system is very sensitive to the laser spot position. The response is determined by transition characteristics like the center, curvature, and length. In this paper, by using the thermal Williams-Comstock model and the microtrack model, the effects of laser spot position on transition characteristics are investigated for both longitudinal and perpendicular recording from a read channel systems perspective. The general trend in their variation is determined and used to explain the resultant change in response. By simulation, we determine the post-Viterbi dominant error events for longitudinal HAMR system. Finally, we apply the well-known MTR codes to mitigate such errors and present their bit-error performance

Analytical Performance of One-Step Majority Logic Decoding of Regular LDPC Codes

In this paper, we present a combinatorial algorithm to calculate the exact bit error rate performance of regular low-density parity check codes under one-step majority logic decoding. Majority logic decoders have regained importance in nano-scale memories due to their resilience to both memory and logic gate failures. This result is an extension of the work of Rudolph on error correction capability of majority-logic decoders.

Joint Message-Passing Symbol-Decoding of LDPC Coded Signals over Partial-Response Channels

We consider the problem of joint detection and decoding of low-density parity-check (LDPC) coded signals over partial response (PR) channels. A method to graphically represent the constraints imposed by the channel and the code on the channel output sequence is introduced. This enables the design of a detector and decoder that estimates a posteriori probabilities of noiseless channel output symbols rather than binary channel inputs. By running the sum-product algorithm (SPA) on this graph, a joint decoder is obtained that is shown to perform significantly better than the turbo-equalizer, at the cost of increased computational complexity.

Bi-Directional Pattern-Dependent Noise Prediction for Heat-Assisted Magnetic Recording With High Jitter Noise

In this paper, we first report the results of an investigation of perpendicular heat-assisted magnetic recording (HAMR) channels containing high jitter noise that is likely to be the case at very high areal densities (e.g., 4 Tb/in2). To model the HAMR channel, we use the thermal Williams-Comstock model and the microtrack model to derive the transition response, without large thermal spot approximation. Further, we propose a novel bi-directional pattern-dependent noise prediction (biPDNP) detector to improve the performance of the HAMR channel under high jitter noise conditions. The biPDNP detector utilizes backward linear prediction in the noise prediction process, as well as the conventional forward linear prediction. At bit error rate of 10-3, biPDNP detector offers 0.6-1.4 dB performance gain over the conventional PDNP detector.

Reverse concatenated watermark codes

In this paper, we introduce an error correction coding scheme, known as reverse concatenated watermark codes for substitution, insertion and deletion channel. This code is obtained by concatenating a distribution transformer with an LDPC code. If the distribution transformer is implemented using a sparsify code then this code is the reverse concatenation of inner and outer codes of the watermark scheme introduced by Davey and Mackay [1]. We show that reverse concatenation significantly decreases decoder complexity and enables an increase in the overall code rate without a prohibitive increase in complexity, although it incurs a loss in performance.

Error and erasure rates for two-dimensional magnetic recording systems

In this paper, we apply a method to view two-dimensional magnetic recording (TDMR) system as a binary error and erasure channel to experimentally determine some bounds on capacity for the TDMR channel. We analyze two different TDMR read-channel models, Voronoi-grain model and random-grain model, to determine the effects of write-errors on the capacity bounds.