Detection and Decoding for 2D Magnetic Recording Channels with 2D Intersymbol Interference

Jiyang Yu, Michael Carosino, Krishnamoorthy Sivakumar, Benjamin J. Belzer, Yiming Chen

Abstract

This paper considers iterative detection and decoding on a concatenated communication channel consisting of a two-dimensional magnetic recording (TDMR) channel modeled by the four-grain rectangular discrete grain model (DGM), followed by a two-dimensional intersymbol interference (2D-ISI) channel. An iterative detection and decoding scheme combines TDMR detection, 2D-ISI detection, and soft-in/soft-out (SISO) channel decoding in a structure with two iteration loops. Simulation results for the concatenated TDMR and 2x2 averaging mask ISI channel with 10 dB SNR show that densities of 0.48 user bits per grain and above can be achieved, corresponding to an areal density of about 9.6 Terabits per square inch.

Introduction

Traditional magnetic recording systems write data on one-dimensional tracks. As storage density increases, magnetic grains on the disk surface become too small to remain stable, hitting a physical limit known as the superparamagnetic limit. Two-Dimensional Magnetic Recording (TDMR) offers a path forward by writing and reading bits across the disk surface in two dimensions. However, TDMR introduces significant challenges: the overwrite property, where random grain sizes cause bits to overwrite previous ones, and 2D Intersymbol Interference (2D-ISI), where the read-head blurs adjacent bits horizontally and vertically during the read process.

System Model

The researchers utilize a "four-grain rectangular discrete grain model" (DGM) to represent the magnetic medium. In this model, bits are mapped to a grid of tiles forming grains of various sizes ($1\times1$, $2\times1$, $1\times2$, or $2\times2$). The output of this grain model is then processed through a 2x2 averaging mask to simulate 2D-ISI, followed by the addition of additive white Gaussian noise (AWGN).

Block diagram of the TDMR system model including the magnetic grain model and 2D ISI channel.

Iterative Receiver Architecture

To combat the combined effects of grain-level overwriting and spatial blurring, the paper proposes a receiver with a double-loop iterative structure. This architecture is designed to decouple the 2D-ISI equalization from the grain-aware detection and the final channel decoding.

  1. Outer Loop: This loop exchanges soft information (Log-Likelihood Ratios or LLRs) between a 2D-ISI estimator and a magnetic grain estimator. The ISI estimator addresses the blurring, while the grain estimator handles the non-linear overwrite effects.
  2. Inner Loop: This loop exchanges LLRs between the magnetic grain estimator and a Serially Concatenated Convolutional Code (SCCC) decoder.
The double-loop iterative receiver architecture showing the exchange of LLRs between the ISI estimator, grain estimator, and channel decoder.

A critical component of this architecture is the introduction of a multiplicative weighting factor ($w \approx 0.26$) for the LLRs exchanged between the ISI and grain estimators. This factor prevents "overconfident" but incorrect LLRs from propagating through the system, which is essential due to the non-linear nature of the grain overwrite model. Standard LLR subtraction used in Turbo codes was found to be less effective than this weighted approach.

Key Results and Insights

The proposed system demonstrates high resilience to 2D-ISI, suffering only a minor rate penalty (7-10%) compared to a clean TDMR channel. Using Extrinsic Information Transfer (EXIT) charts, the authors optimized the iteration counts and weighting factors. They also discovered that the Generalized Extreme Value (GEV) distribution provides an excellent fit for the LLR histograms produced by the detector, aiding in more accurate performance modeling during the optimization phase.

The system achieved a storage density of 0.48 user bits per grain. Assuming a standard media density of 20 Teragrains/in², this translates to an areal density of 9.6 Terabits/in². This performance represents a significant improvement over previous state-of-the-art methods, outperforming them by up to 5.5% in code rate within realistic grain probability ranges.

Conclusion

This work demonstrates that iterative detection and decoding can effectively manage the complex distortions found in TDMR systems. While the 4-grain rectangular model provides a solid theoretical foundation, future research should explore more realistic "sea-of-grains" models and modern coding schemes like Low-Density Parity-Check (LDPC) codes to further close the gap to theoretical channel capacity.

Summary updated

Abstract

This paper considers iterative detection and decoding on the concatenated communication channel consisting of a two-dimensional magnetic recording (TDMR) channel modeled by the four-grain rectangular discrete grain model (DGM) proposed by Kavcic et. al., followed by a two-dimensional intersymbol interference (2D-ISI) channel modeled by linear convolution of the DGM model's output with a finite-extent 2D blurring mask followed by addition of white Gaussian noise. An iterative detection and decoding scheme combines TDMR detection, 2D-ISI detection, and soft-in/soft-out (SISO) channel decoding in a structure with two iteration loops. In the first loop, the 2D-ISI channel detector exchanges log-likelihood ratios (LLRs) with the TDMR detector. In the second loop, the TDMR detector exchanges LLRs with a serially concatenated convolutional code (SCCC) decoder. Simulation results for the concatenated TDMR and 2 x 2 averaging mask ISI channel with 10 dB SNR show that densities of 0.48 user bits per grain and above can be achieved, corresponding to an areal density of about 9.6 Terabits per square inch, over the entire range of grain probabilities in the TDMR model.
BibTeX
@misc{yu2014detection,
  author = {Jiyang Yu and Michael Carosino and Krishnamoorthy Sivakumar and Benjamin J. Belzer and Yiming Chen},
  title = {{Detection and Decoding for 2D Magnetic Recording Channels with 2D Intersymbol Interference}},
  year = {2014},
  eprint = {1401.4506},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1401.4506},
}
Figure