Construction of Quantum Rank-Metric Codes Using Hermitian Orthogonality

TL;DR

This paper introduces a new method for constructing quantum rank-metric codes using Hermitian orthogonality, enabling codes for even-sized stacked quantum memories and improving error correction.

Contribution

It proposes a novel construction framework for quantum Gabidulin codes leveraging Hermitian self-orthogonality, overcoming previous shape restrictions.

Findings

Approximately doubled the minimum rank distance ratio

Eliminated the odd-size restriction for stacked memories

Enhanced error correction and design flexibility

Abstract

Stacked quantum memory is an architecture in which multiple layers of qubits are stacked. Quantum rank-metric codes are effective for error correction in stacked quantum memories. However, the previously proposed quantum Gabidulin codes based on the CSS construction had a problem: due to algebraic constraints, the applicable memory layouts were strictly limited to square shapes of odd length. In this paper, we first propose a framework for constructing quantum rank-metric codes from classical linear codes with symplectic self-orthogonality. Building upon this, we propose a new construction method for quantum Gabidulin codes by combining the Hermitian self-orthogonality of classical Gabidulin codes--utilizing the self-dual basis that exists when the extension degree of the finite field is even--with the quantum code construction method using Hermitian orthogonality by Matsumoto and Uyematsu. The proposed method succeeds in approximately doubling the ratio of the minimum rank distance to the number of physical qubits while maintaining the code rate. Furthermore, it eliminates the restriction of the conventional method that requires the number of cells and layers of the stacked memory to be odd, realizing the construction of quantum rank-metric codes applicable to memories with an even number of cells and layers. This construction improves the relative error correction capability of the stacked quantum memory architecture and increases the degree of freedom in design while preserving the code rate.