NCR-9415374, U. of Hawaii, Manoa, "Soft-Decision Decoding, Trellis Structure and Coded Modulation", PI-Lin: As the demand for data integrity increases, coding for error control becomes increasingly important in data transmission systems. It has become an integral part in almost every data communication system design. Today very sophisticated error control coding schemes are being used in a broad range of data communication systems to achieve reliable data transmission. The objective of this research project is to investigate several problems in coding which are both theoretically and practically important. The research includes three inter-related areas: (1) soft-decision decoding of linear block codes; (2) trellis structure of linear block codes; and (3) coded modulation. In the first research area, the research will devise efficient soft-decision decoding schemes for linear block codes to achieve optimum or near optimum error performance with reduced decoding complexity. The approaches to be taken include: (1) hard-decision decoding based on certain reliability measures of the received symbols; (2) iterative reprocessing of the hard-decision decoded word until optimum or near optimum error performance is achieved; (3) searching for the maximum likelihood solution through a purged trellis diagram for the code to be decoded; and (4) multistage soft-decision decoding based on code decomposition. In the second research area, the trellis structure of linear block codes is investigated in terms of the state complexity, the branch complexity, the state connectivity and parallel structure of minimal trellis diagrams. For codes with simple trellis structure, the Viterbi algorithm can be applied to achieve maximum likelihood decoding with reduced computational complexity. Other problems to be investigated in this area include: (1) construction of minimal trellis diagrams for known codes; (2) finding specific permutations on the bit positions of codes that minimize or reduce the structural complexity of code trellises; and (3) state and branch labeling. In the third research area, the investigation of coded modulation will be continued for bandwidth efficient communications. Modulation codes for both AWGN and fading channels will be constructed and analyzed. Emphasis is placed in constructing modulation codes or coded modulation systems to achieve high performance with reduced decoding complexity. One approach to be taken is to combine coded modulation and product coding technique. Unequal error protection modulation codes will also be investigated. The research is intended to advance the state of knowledge in coding theory and error control techniques. Fruitful results from this study will have significant impact on coding research and the design of efficient error control systems for reliable data communications. ***************************************************************************** Aubrey M. Bush Program Director, Acting Deputy Divison Director Division of Networking and Communications Research and Infrastructure National Science Foundation
