東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

Graphical models for iterative data ...

Thiennviboon, Phunsak.

Linked to FindBook

Google Book

Amazon

博客來

Graphical models for iterative data detection.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Graphical models for iterative data detection./
Author:	Thiennviboon, Phunsak.
Description:	223 p.
Notes:	Source: Dissertation Abstracts International, Volume: 64-06, Section: B, page: 2844.
Contained By:	Dissertation Abstracts International64-06B.
Subject:	Engineering, Electronics and Electrical. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3094422
ISBN:	0496421231

Graphical models for iterative data detection.
Thiennviboon, Phunsak.

Graphical models for iterative data detection. - 223 p.

Source: Dissertation Abstracts International, Volume: 64-06, Section: B, page: 2844.

Thesis (Ph.D.)--University of Southern California, 2002.

A graphical modelling is the art of system modelling originally developed in many areas, especially for inference problems in statistics and expert systems. Recently, the connection with decoding/detection algorithms in communications was discovered. Although this method is becoming prevalent tool in the coding community, it has received relatively little attention in many practical communication problems.

ISBN: 0496421231Subjects--Topical Terms:

626636
Engineering, Electronics and Electrical.

Graphical models for iterative data detection.
LDR:03395nmm 2200313 4500 001 1841532
005 20050915142049.5
008 130614s2002 eng d
020 $a 0496421231
035 $a (UnM)AAI3094422
035 $a AAI3094422
040 $a UnM $c UnM
100 1 $a Thiennviboon, Phunsak. $3 1929825
245 1 0 $a Graphical models for iterative data detection.
300 $a 223 p.
500 $a Source: Dissertation Abstracts International, Volume: 64-06, Section: B, page: 2844.
500 $a Adviser: Keith M. Chugg.
502 $a Thesis (Ph.D.)--University of Southern California, 2002.
520 $a A graphical modelling is the art of system modelling originally developed in many areas, especially for inference problems in statistics and expert systems. Recently, the connection with decoding/detection algorithms in communications was discovered. Although this method is becoming prevalent tool in the coding community, it has received relatively little attention in many practical communication problems.
520 $a In this thesis, graphical models are represented by junction graphs together with various types of message-passing algorithms (MPAs). The general sufficient conditions for well-defined marginalizing-combining operations of MPAs are proposed to account both continuous and discrete variables and three practical problems involving iterative data detection are considered. First, a full binary tree is introduced to the soft-input/soft-output (SISO) module in order to obtain a low-latency efficient parallel-processing implementation. In other words, the forward-backward-tree (FBT) SISO algorithm provides O(log 2 N) latency as opposed to O(N) latency from the standard forward-backward (FB) algorithm for the SISO module.
520 $a Next, the general two-dimensional data detection is considered and is proven to be NP hard. Three types of loopy graphical models are investigated including the simple input-output (SIO) model, the serially concatenated (SC) model, and the novel grid model, along with standard and modified MPAs. Simulation results for the page-oriented optical memory system show that these iterative algorithms can provide near-optimal performance while the best performance can be obtained through the SC or grid models. Moreover, a reduced-complexity grid algorithm is proposed and applied to digital image halftoning with complexity and image quality comparable to the best techniques in the literature.
520 $a Finally, the general structure of trellis-based reduced-state (RS) FB algorithms is proposed. This structured is described from ad hoc intuitive approach and then systematically developed based on the loopy overstructured graph (OSG) and a novel modified MPA. While maintaining the optimality in the absence of state reduction, the modified MPA provides a broad class of existing and novel practical RS-FB algorithms through applications of reduced-state techniques. Extensive numerical simulations are considered and it is demonstrated that in many cases the performance of RS-FB algorithms is close to that of the corresponding full-complexity algorithms.
590 $a School code: 0208.
650 4 $a Engineering, Electronics and Electrical. $3 626636
650 4 $a Computer Science. $3 626642
690 $a 0544
690 $a 0984
710 2 0 $a University of Southern California. $3 700129
773 0 $t Dissertation Abstracts International $g 64-06B.
790 1 0 $a Chugg, Keith M., $e advisor
790 $a 0208
791 $a Ph.D.
792 $a 2002
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3094422