A procedure for designing FDTS/DF as a high dimensional signal space detector is presented. The procedure is applied to the three dimensional case to illustrate the resulting detector structure. An equalization and code constraint reduce the number of boundaries and to eliminate the multipliers from the general case. Simulation results show this new channel outperforms EPRML for user densities of 2.25 and higher.

Maximum transition run (MTR) codes provide significant minimum distance gains when used with sequence detectors operating at high linear densities. A method for reducing the RLL k constraint associated with MTR block codes is presented. A block decodable, rate 4/5 MTR code with k=4 illustrates the technique. This reduction of k is combined with sliding-block decoding to develop a 97.8% efficient rate 6/7 MTR code with k=8.

This paper addresses the optimality of the fixed delay tree search with decision feedback (FDTSIDF) detector for data without a minimum run length constraint. The evaluation is performed by comparing the minimum distance for FDTSIDF with the minimum distance associated with a maximum likelihood sequence detector (MLSD). Using this distance information, potential improvements through coding areaddressed.An alternative realization to the direct implementation of FDTS/DF is considered by examining the detection process in the context of a signal space.

A b s t r a c t -A new code is presented which improves the minimum distanceproperties of sequence detectors operating at high linear densities. This code, which is called the maximum transition run code, eliminates data patterns producing three or more consecutive transitions while imposing the usual k-constraint necessary for timing recovery. The code possesses the simikr distance-gaining property of the (1,k) code, but can be implemented with considerably higher rates. Bit error rate simulations on fixed delay tree search with decision feedback and high order partial response maximum likelihood detectors confirm large coding gains over the conventional (0,k) code.

No abstract