Iterative row-column soft-decision feedback algorithm using joint extrinsic information for two-dimensional intersymbol interference

Chen, Yiming; Belzer, Benjamin J.; Sivakumar, Krishnamoorthy

doi:10.1109/ciss.2010.5464706

Cited by 13 publications

(9 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…2. The receiver consists of the joint row-column 2D-ISI equalizer [13] and IRA decoder. After iterating between row and column detectors, the 2D-ISI equalizer sends binary LLR estimates of the coded bitsŷ to the IRA decoder after subtracting out the corresponding input LLRs received from the IRA decoder at the start of the current iteration.…”

Section: B Signal Processing System For Tdmr Modelmentioning

confidence: 99%

EXIT chart based IRA code design for TDMR

Mehrnoush

Belzer

Sivakumar

et al. 2016

2016 Annual Conference on Information Science and Systems (CISS)

Self Cite

View full text Add to dashboard Cite

This paper presents an extrinsic information transfer (EXIT) chart based irregular repeat-accumulate (IRA) code design technique for a two-dimensional magnetic recording (TDMR) turbo-equalizer that employs a Voronoi magnetic grain model. The channel model also includes two-dimensional intersymbol interference (2D-ISI) and additive white Gaussian noise (AWGN). At high bit densities (e.g., between 1 and 3 magnetic grains per coded bit (GPB)), occasionally a bit will not be written on any grain, and hence will effectively be "overwritten" (or erased) by bits on surrounding grains. The proposed code design takes into account the statistics of the TDMR channel to decrease overwrite effects. The proposed receiver uses a 2D-ISI BCJR equalizer and IRA decoder in a turbo-equalization approach. To design the IRA code, we find the experimental EXIT chart curves of the check node decoder (CND) and the combination of the variable node decoder (VND) with the 2D-ISI equalizer. We fit the VND and CND EXIT chart curves to find the IRA code's optimized variable node degree distribution for the TDMR channel. Simulation results show that the IRA codes optimized for the TDMR Voronoi grain model achieve up to a 3.3% density increase in user-bits/grain (U/G) compared to IRA codes designed for AWGN channels. At 1.2 GPB, the designed IRA codes achieve densities as high as 0.455 U/G, corresponding to an areal density of about 4.6 Terabits/in 2 on typical magnetic hard disks with 10 Teragrains/in 2 ; this is nearly a factor of 5 better than the best commercially available systems.

show abstract

Section: B Signal Processing System For Tdmr Modelmentioning

confidence: 99%

EXIT chart based IRA code design for TDMR

Mehrnoush

Belzer

Sivakumar

et al. 2016

2016 Annual Conference on Information Science and Systems (CISS)

Self Cite

View full text Add to dashboard Cite

show abstract

“…3. The receiver consists of the joint row-column 2D-ISI equalizer [10] and IRA decoder. After iterating between row and column detectors, the 2D-ISI equalizer sends binary LLR estimates of the coded bitsŷ to the IRA decoder after subtracting out the corresponding input LLRs received from the IRA decoder at the start of the current iteration.…”

Section: A Transmitter and Receiver Architecturesmentioning

confidence: 99%

“…4 from [10], they must be marginalized to binary LLRs before passing them to the IRA decoder. The binary LLR for position k (the bit position of i 0 in the 2D-ISI row detector's state/input block shown in Fig.…”

Section: A Transmitter and Receiver Architecturesmentioning

confidence: 99%

“…Fig. 4: State, input, and feedback pixels for the BCJR rowdetector of the 2D-ISI equalizer (from [10]). …”

Section: A Transmitter and Receiver Architecturesmentioning

confidence: 99%

“…The present paper follows the same general approach as [2], except that the MMSE equalizer is replaced with a higher performance iterative row-column BCJR equalizer from [10], the LDPC code is replaced by a higher performing irregular repeat-accumulate (IRA) code [11], and the perturbed-bitcenter Voronoi model is replaced by the Voronoi grain model of [3], which starts with randomly distributed grain nuclei that are completely independent of the coded-data-bit center locations. We use the same 2D-ISI impulse response function as [2] to facilitate comparison between the two approaches.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Signal processing for two dimensional magnetic recording using Voronoi model averaged statistics

Mehrnoush

Belzer

Sivakumar

et al. 2015

2015 49th Annual Conference on Information Sciences and Systems (CISS)

Self Cite

View full text Add to dashboard Cite

This paper considers a signal processing system for two-dimensional magnetic recording (TDMR) employing a random Voronoi grain model. The channel model also includes two-dimensional intersymbol interference (2D-ISI) and additive white Gaussian noise. The system uses a 2D-ISI BCJR detector and irregular repeat-accumulate (IRA) code decoder in a turboequalization approach. In order to transfer soft information to the IRA decoder a mapping function based on the 2D-ISI detector soft information statistics is used. Simulations employing the perturbed-bit-centers Voronoi grain model proposed in a previous paper by Hwang et al. show that the proposed system achieves a 6.5% increase in user bits/grain (U/G) and an 11.7 dB SNR gain compared to Hwang et al. Simulation results also indicate that our random Voronoi model is harder to equalize than the Hwang Voronoi model. For the random Voronoi model considered in this paper, the proposed system achieves a density of 0.4422 U/G, corresponding to an areal density of about 8.8 Terabits/in 2 on typical magnetic hard disks; this is nearly an order of magnitude better than the best commercially available systems.

show abstract