Constant Euclidean-Norm Perfect Gaussian Integer Sequences

Chong‐Dao Lee, Yan-Haw Chen, C. L. Philip Chen, Yaotsu ChangIEICE 2026 · DOI

Introduction

In the realm of digital communications, particularly within systems like Orthogonal Frequency Division Multiplexing (OFDM) and Code Division Multiple Access (CDMA), the design of complex-valued sequences is critical for managing signal power and ensuring reliable synchronization. A "Perfect" sequence is defined by its periodic autocorrelation property: it must be zero for all non-zero shifts. This mathematical uniqueness allows receivers to distinguish a sequence from its shifted versions, significantly reducing interference.

However, a persistent challenge in these systems is the Peak-to-Average Power Ratio (PAPR). To optimize power amplifier efficiency and minimize signal distortion, engineers aim for sequences where every element has the same magnitude, or Euclidean norm. This state is characterized by an energy efficiency ($\eta_S$) of 1.0. While achieving this is straightforward for odd-length sequences, even-length sequences have historically been limited to efficiencies around 0.75. This paper, published in IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences, introduces a novel framework to construct even-length Perfect Gaussian Integer Sequences (PGISs) that finally reach the 1.0 efficiency threshold.

Construction Methodology: The Pythagorean Approach

The researchers propose a construction method that leverages the combination and transformation of "base sequences." The fundamental innovation is the application of the Pythagorean theorem ($a^2 + b^2 = c^2$) to the selection of Gaussian integers.

By treating the real and imaginary parts of a Gaussian integer as the legs of a right triangle, the authors can generate a variety of distinct elements that all share the same hypotenuse—and thus the same Euclidean norm. For instance, using the Pythagorean triple (3, 4, 5), Gaussian integers such as $3+4j$, $4+3j$, $5$, and $-4+3j$ all possess a norm of 5.

Visual representation of Gaussian integers with constant Euclidean norms on the complex plane.

The construction process involves three primary steps:

  1. Base Sequence Selection: Starting with known perfect integer sequences (e.g., $[1, -1, -1, -1]$ for length 4).
  2. Cyclic Shifting and Interleaving: Applying specific cyclic shifts and combining real and imaginary components to preserve the perfect autocorrelation property.
  3. Pythagorean Mapping: Mapping sequence elements to Gaussian integers derived from Pythagorean triples to ensure a constant magnitude across the entire sequence.

The paper provides rigorous theoretical guarantees (Theorems 1-4) for sequences of lengths 4, 8, and 16.

Key Insights and Results

The most significant breakthrough of this work is breaking the 0.75 energy efficiency barrier for even-length sequences. By hitting the 1.0 mark, these sequences offer maximum energy efficiency for hardware implementations.

Furthermore, the proposed method provides a much higher "degree of freedom" compared to previous state-of-the-art techniques, such as the Hu-Wang-Li sequences. The "degree" refers to the number of unique Gaussian integers present in a sequence. For a length-16 sequence, the new method can achieve a degree of up to 11, whereas previous methods were significantly more restricted. This variety gives engineers greater flexibility when designing sequences for specific channel conditions or hardware constraints.

The flexibility of the construction also allows the Euclidean norm to be either an integer (via Pythagorean triples) or an irrational number (e.g., $\sqrt{2}$), depending on the initial parameters chosen for the base sequences.

Conclusion and Future Directions

While the current study focuses on short sequence lengths (4, 8, and 16), these are highly applicable in specific low-latency or low-complexity communication protocols. The researchers have demonstrated that the previous efficiency limits were not fundamental mathematical ceilings but rather limitations of earlier construction methodologies.

Future work will focus on extending these constant Euclidean-norm properties to longer sequences. Scaling the construction while maintaining the "perfect" autocorrelation requirement remains a primary objective for the next phase of research in high-efficiency signal design.

Summary updated

Abstract

This letter presents a new class of perfect Gaussian integer sequences (PGISs), called Constant Euclidean-norm PGISs, where every Gaussian integer has the same Euclidean norm. Several proposed construction methods are based on the combination of different base sequences. With the aid of Pythagorean theorem, distinct elements appearing in the base sequences result in many new families of constant Euclidean-norm PGISs with degree of freedom. Both theoretical guarantees and numerical examples of the newly proposed construction methods are provided as well.
BibTeX
@inproceedings{lee2026constant,
  author = {Chong‐Dao Lee and Yan-Haw Chen and C. L. Philip Chen and Yaotsu CHANG},
  title = {{Constant Euclidean-Norm Perfect Gaussian Integer Sequences}},
  booktitle = {IEICE},
  year = {2026},
  doi = {10.1587/transfun.2025eal2108},
}
Figure