[Paper Review] On the Duality of Erasures and Defects

Yongjune Kim, B. V. K. Vijaya Kumar|arXiv (Cornell University)|Mar 7, 2014

Error Correcting Code Techniques16 references3 citations

TL;DR

This paper establishes a duality between erasures and defects in linear coding theory by analyzing the solvability of systems defined by generator matrices and received vectors. It proves that a solution exists if and only if the rank of the augmented matrix equals the rank of the generator submatrix, with key results showing that when the rank is full, the probability of no errors (M=0) is zero, and when the rank is deficient, solvability depends on the structure of the last j elements of the received vector.

ABSTRACT

In this paper, the duality of erasures and defects will be investigated by comparing the binary erasure channel (BEC) and the binary defect channel (BDC). The duality holds for channel capacities, capacity achieving schemes, minimum distances, and upper bounds on the probability of failure to retrieve the original message. Also, the binary defect and erasure channel (BDEC) will be introduced by combining the properties of the BEC and the BDC. It will be shown that the capacity of the BDEC can be achieved by the coding scheme that combines the encoding for the defects and the decoding for the erasures. This coding scheme for the BDEC has two separate redundancy parts for correcting erasures and masking defects. Thus, we will investigate the problem of redundancy allocation between these two parts.

Motivation & Objective

• To investigate the mathematical conditions under which error patterns in linear codes can be uniquely decoded.
• To establish a formal duality between erasures and defects in coding systems using matrix rank analysis.
• To determine the solvability of decoding equations based on the rank of submatrices derived from generator matrices and received vectors.
• To quantify the probability of successful decoding (M=0) under different rank conditions of the system matrix.

Proposed method

• Analyzes the solvability of a system of linear equations defined by a generator matrix G₀ᵁ and a received vector bᵁ.
• Uses the rank condition rank(G₀ᵁ) = rank(G₀ᵁ | bᵁ) as a necessary and sufficient criterion for the existence of at least one solution.
• Applies row-reduced echelon form to identify zero rows in G₀ᵁ when its rank is less than u.
• Examines the last j elements of bᵁ to determine whether the system remains solvable when the rank of G₀ᵁ is reduced by j.
• Derives the probability P(M=0 | |U|=u) based on the rank of G₀ᵁ, showing it is zero when rank(G₀ᵁ) = u.
• Considers the case where rank(G₀ᵁ) = u−j for 1 ≤ j ≤ u, and investigates the implications for solvability and error correction.

Experimental results

Research questions

• RQ1 Under what conditions does the system of equations defined by the generator matrix and received vector have at least one solution?
• RQ2 How does the rank of the generator submatrix G₀ᵁ relate to the solvability of the decoding problem?
• RQ3 What is the relationship between the rank deficiency of G₀ᵁ and the structure of the received vector bᵁ?
• RQ4 How does the probability of no errors (M=0) depend on the rank of G₀ᵁ when |U|=u?
• RQ5 What is the role of the last j elements of bᵁ in determining the solvability of the system when rank(G₀ᵁ) = u−j?

Key findings

• A solution to the system exists if and only if the rank of the augmented matrix (G₀ᵁ | bᵁ) equals the rank of G₀ᵁ.
• When rank(G₀ᵁ) = u, the system is guaranteed to have at least one solution, leading to P(M=0 | |U|=u) = 0.
• If rank(G₀ᵁ) = u−j for 1 ≤ j ≤ u, the last j rows of the row-reduced echelon form of G₀ᵁ are zero vectors.
• The solvability of the system depends critically on the last j elements of the vector bᵁ when the rank is deficient.
• The probability of no errors (M=0) is zero when the generator submatrix has full rank u.