On Tanner Graphs of Lattices and Codes - Information Theory, 1998. Proceedings. 1998 IEEE International Symposium on
نویسندگان
چکیده
The problem of finding a low-complexity Tanner graph for a general lattice A is studied. The problem is divided into two subproblems: 1) Finding an orthogonal sublattice A‘ of A which minimizes the complexity of the label code of the quotient group A/A’. 2) Constructing a simple Tanner graph for the label code obtained in part 1. The proposed approach for solving subproblem 2 can also be applied to any abelian group block code with arbitrary finite alphabets at different coordinates. The results of this work are useful in finding low-complexity graph representation of lattices and codes, which consequently results in efficient graph-based decoding algorithms. I. SUMMARY OF RESULTS The graph representation of codes and the corresponding decoding algorithms, initiated by the work of Gallager on lowdensity parity-check codes [2] and later generalized by Tanner [3] to codes defined by general bipartite graphs, have continued to be an active area of research particularly in the past few years (see, e.g., [SI and the references therein). The construction of Tanner graphs for linear block codes is well-known and is based on using a parity-check matrix of the code [3]. A Tanner graph construction for lattices, using a method different from the one discussed in this paper, was briefly sketched in [5]. In this work, we develop Tanner graph constructions for abelian group block codes, an important application of which is to represent the label code of a lattice [l] in a given coordinate system. Our study shows that to obtain a lowcomplexity Tanner graph for an n-dimensional (n-D) lattice A, it is reasonable to divide the problem into two subproblems: 1) finding a set of 1-D Orthogonal subspaces {Wi};’,, called graph coordinate system, which minimizes the sizes gi of the label groups Gi = Pw,(A)/Aw,, where Pw,(A) and Aw, denote the projection and cross section of A on Wi, 2) deriving a simple Tanner graph for the corresponding label code of A/A’, where A’ is the orthogonal sublattice of A in {Wi}T=l. In the following, we first address subproblem 2. To simplify the presentation of the label code, we consider its isomorphic group defined over the alphabet d = Zcl x Z, x * * * x Z, , where Z, = {0,1,. .. ,g; 1). This is denoted by G(A). The dual G’ of G is also defined over A. It consists of those codewords c* = (c:, . . . , cz) E A such that E:=’=, cfci/gi E Z, Vc = (cl,. . . , c,) E G, where multiplications and divisions are performed in the field of real numbers. It can be seen that G’ is also an abelian group with order IG’l = ldl/lGl. The label code G is therefore fully described by the following set of check equations:
منابع مشابه
Tanner graphs for group block codes and lattices: Construction and complexity
We develop a Tanner graph (TG) construction for an Abelian group block code with arbitrary alphabets at different coordinates, an important application of which is the representation of the label code of a lattice. The construction is based on the modular linear constraints imposed on the code symbols by a set of generators for the dual code . As a necessary step toward the construction of a TG...
متن کاملOptimal Sequential Vector Quantization of Markov Sources - Information Theory, 1998. Proceedings. 1998 IEEE International Symposium on
The problem of optimal sequential vector quantization of Markov sources is cast as a stochastic control problem wi th partial observations and constraints, leading to useful existence results for optimal codes and their characterizations.
متن کاملConvolutional Code Constructions Resulting in Maximal or near Maximal Free Distance - Information Theory, 1998. Proceedings. 1998 IEEE International Symposium on
In this paper we discuss an upper bound on the free distance for a rate k / n convolutional code with complexity b . Using th i s bound we in t roduce the notion of a MDS convolutional code. We also give an algebraic way of cons t ruc t ing b inary codes of rate 1/2 and large complexity. The obta ined distances compare favorably to the distances found by computer searches and probabilistic meth...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2004