Coding, Communications and Information Theory Group
- Coding for high density magnetic and optical recording
- Turbo codes and error control coding
- Source coding and data compression
Shannon’s 1948 paper laid the ground work for much of what is done in communications today: data compression, error control coding, modulation, and even networks. We are interested in the application of communications and information theory to real world channels and systems. Currently are efforts are oriented to high density magnetic and optical recording.
NONBINARY OPTICAL RECORDING
Related Sites For a list of other companies and universities doing research in this area check out this site.
Introduction We have developed coding and signal processing methods to increase storage density on a optical disk. Our primary efforts are in the area of M-ary (or nonbinary) recording channels. A number of optical media have been proposed for nonbinary recording: electron trapping optical memory (by ETOM Technologies), pit depth modulation (by Calimetrics) magnetic multivalued (MMV) magneto-optic media (by Hitachi Maxell), nonbinary phase change materials (several companies).
Modulation Codes. We have developed numerous codes are of two general types: runlength limited (RLL) M-ary codes and partial response (PR) codes. A list of recent publications address in this area is here. We hold 14 US patents and have 8 patents pending in this area.
Fundamental Limits . We have also established fundamental (information theoretic) limits in optical recording. Chief among them is a result on the rationality of the capacity of M-ary RLL constrained channels. We have shown that the capacity is irrational except in a few special cases. In those cases where the capacity is rational, we have designed provably minimum complexity, capacity-achievingcodes. Again, we have published numerous papers listed here.
Support Our work in this area has been generously supported by Optex Communication (now ETOM Technologies), The National Institute of Standards and Technology (NIST) Advanced Technology Program, The National Science Foundation PECASE award , and NSF Grant Opportunities for Aacademic Liason with Industry (GOALI) award . We are recently started a collaboration with three companies, Calimetrics , Polaroid , and Energy Conversion Devices , and are currently seeking funding to use our joint expertise in nonbinary recording media to develop new signal processing and and coding approaches.
CODING FOR HIGH DENSITY MAGNETIC RECORDING
Turbo codes for high density magnetic recording. . We applied turbo codes to partial response digital magnetic recording and showed that high rate turbo codes (4/5, 8/9 16/17) providesignificant coding gain over conventional PR4 and EPR4. . For several recent conference papers checkout CISS98_turbo_p r4.ps Asilomar99.ps, Intermag99.ps. We are presently addressing the following issues: turbo codes for partial response channels, turbo codes with runlength constraints, reduced complesity (stack) decoding of turbo codes, hardware implementation of turbo codes, low density parity check codes and block turbo codes.
Coding for constrained channels . We have developed several types of coding methods for binary channels. First, a new constraint known as the “one pairs constraint” is used to improve timing recovery and eliminate quasi-catastrophic error events in PR magnetic recording. A paper describing this was presented at Globecom97 Globecom97.ps, and will be appearing in the IEEE Transactions on Magnetics in May 1999. We are also investigating very high rate (>99.9% efficient) constrained codes. The codes are constructed using permutation codes and a paper fully describing this will appear in the May 1999 issue of the IEEE Transactions on Information Theory.