Abstract
Trellises provide a graphical representation for the row space of a matrix. The product construction of Kschischang and Sorokine builds minimal conventional trellises from matrices in minimal span form. Koetter and Vardy showed that minimal tail-biting trellises can be obtained by applying the product construction to submatrices of a characteristic matrix. We introduce the unique reduced minimal span form of a matrix and we obtain an expression for the unique reduced characteristic matrix. Among new properties of characteristic matrices we prove that characteristic matrices are in duality if and only if they have orthogonal column spaces, and that the transpose of a characteristic matrix is again a characteristic matrix if and only if the characteristic matrix is reduced. These properties have clear interpretations for the unwrapped unit memory convolutional code of a tail-biting trellis, they explain the duality for the class of Koetter and Vardy trellises, and they give a natural relation between the characteristic matrix based Koetter–Vardy construction and the displacement matrix based Nori–Shankar construction. For a pair of reduced characteristic matrices in duality, one is lexicographically first in a forward direction and the other is lexicographically first in the reverse direction. This confirms a conjecture by Koetter and Vardy after taking into account the different directions for the lexicographical ordering.
Similar content being viewed by others
Notes
\(k \not \in (i,j]\) if the ordered triple i, j, k is even, i.e., \(i< j < k,\)\(k< i < j\) or \(j< k < i.\)
But should not be confused with the different decomposition \(PA=LU.\)
References
Bahl L.R., Cocke J., Jelinek F., Raviv J.: Optimal decoding of linear codes for minimizing symbol error rate. IEEE Trans. Inf. Theory IT–20, 284–287 (1974).
Calderbank A.R., David Forney Jr. G., Vardy A.: Minimal tail-biting trellises: the Golay code and more. IEEE Trans. Inf. Theory 45(5), 1435–1455 (1999).
David Forney Jr. G.: Coset codes-II: binary lattices and related codes. IEEE Trans. Inf. Theory 34(5), 1152–1187 (1988).
David Forney Jr. G.: Minimal realizations of linear systems: the “shortest basis” approach. IEEE Trans. Inf. Theory 57(2), 726–737 (2011).
Duursma I.M.: Two-point coordinate rings for GK-curves. IEEE Trans. Inf. Theory 57(2), 593–600 (2011).
Duursma I.M., Park S.: Delta sets for divisors supported in two points. Finite Fields Appl. 18(5), 865–885 (2012).
Duursma I., Kirov R., Park S.: Distance bounds for algebraic geometric codes. J. Pure Appl. Algebra 215(8), 1863–1878 (2011).
Gluesing-Luerssen H., Weaver E.A.: Linear tail-biting trellises: characteristic generators and the BCJR-construction. IEEE Trans. Inf. Theory 57(2), 738–751 (2011).
Gluesing-Luerssen H., Weaver E.A.: Characteristic generators and dualization for tail-biting trellises. IEEE Trans. Inf. Theory 57(11), 7418–7430 (2011).
Koetter R., Vardy A.: The structure of tail-biting trellises: minimality and basic principles. IEEE Trans. Inf. Theory 49(9), 2081–2105 (2003).
Kschischang F.R., Sorokine V.: On the trellis structure of block codes. IEEE Trans. Inf. Theory 41(6, Part 2), 1924–1937 (1995).
McEliece R.J.: On the BCJR trellis for linear block codes. IEEE Trans. Inf. Theory 42(4), 1072–1092 (1996).
Nori A.V., Shankar P.: Unifying views of tail-biting trellis constructions for linear block codes. IEEE Trans. Inf. Theory 52(10), 4431–4443 (2006).
Springer T.A.: Linear Algebraic Groups. Modern Birkhäuser Classics, 2nd edn. Birkhäuser Boston, Inc., Boston (2009).
Vardy A.: Trellis structure of codes. In: Handbook of Coding Theory, vol. I, II, pp. 1989–2117. North-Holland, Amsterdam (1998).
Yerëmin A.Y.U., Kolotilina LYu: The Bruhat decomposition and the solution of systems of linear algebraic equations. Sov. J. Numer. Anal. Math. Model. 2(6), 421–436 (1987).
Acknowledgements
This work was supported in part by the Simons Foundation under Grant 280107 and in part by NSF under Grant CCF-1619189.
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by C. Mitchell.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendix A: A dual pair of reduced characteristic matrices
Reduced characteristic matrices X and Y for a pair of binary orthogonal matrices G and H (Figs. 3, 4, 5).
Label codes for KV-trellises constructed from X and Y using the span based BCJR construction.
Label codes for the dual trellises defined with rows 1, 3, 5 of X and rows 2, 4 of Y.
Appendix B: A reduced characteristic matrix for the Golay code
The binary Golay code of length 24 is generated by the following twelve row vectors over \(GF(4) = \{ 0,1,a,b \}\) after the concatenation \(0 \mapsto 00, 1 \mapsto 11, a \mapsto 01, b \mapsto 10.\)
Every even row is a scalar multiple of the preceding odd row and every odd row is obtained with an obvious symmetry from the preceding even row. If we apply the concatenation in combination with a transposition of symbols \(4k+2\) and \(4k+4,\) i.e.,
then the result is a generator matrix G for the Golay code with optimal minimal span length \(108 = 12 \cdot 9.\) The generator matrix is of the form (1) in [2] and agrees with the generator matrix in (4) of that paper after a permutation \((1\;2)(7\;8)(9\;10)(15\;16)(17\;18)(23\;24).\) The reduced characteristic matrix X for G is of the form
with (A|B|C) equal to
The odd rows have span length 9 and the even rows have span length 15.
Appendix C: Non-attacking rooks for function fields
Minimal spans for characteristic generators have an analogue in discrepancies for function fields.
Given a field K of algebraic functions and two rational points (places), a function in K has span (i, j) if it has a pole of order j at the first point, no other poles, and a zero of order i at the second point. For each \(j \in \mathbb {Z},\) let \(\sigma (j) \in \mathbb {Z}\) be maximal such that there exists a function with span (i, j). This defines a pair \((\sigma (j),j).\) The pairs fill the plane with an infinite set of non-attacking rooks. Let \(n > 0\) be minimal such that there exists a function with \((i,j)=(n,n).\) Then the pattern is periodic with period n.
For the function field F(x, y), defined with equation \(y^8+y = x^{10}+x^3\) over the field of eight elements, \(n=13.\) There exist functions with minimal spans [6]
Figure 6 pictures the minimal spans as non-attacking rooks on a chess board of size 13 with labeling \(\{ 0, 1, \ldots , 11, 12 \} \times \{ 1, 2, \ldots , 12, 0 \}\) modulo 13.
Rights and permissions
About this article
Cite this article
Duursma, I.M. Matrix theory for minimal trellises. Des. Codes Cryptogr. 87, 2507–2536 (2019). https://doi.org/10.1007/s10623-019-00627-8
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10623-019-00627-8
Keywords
- Linear block code
- Minimal trellis
- Tail-biting trellis
- Characteristic generator
- Koetter and Vardy conjecture