24 April 2024 in Journal of Lightwave Technology
The emergence of advanced technologies has spurred the development of high-capacity, long-distance, and high-speed coherent optical communication systems. However, Chromatic Dispersion (CD) is the major challenge of coherent optical communication leading to high power consumption at the receiver which impedes the adoption of the technology. The existing systems adopt a high-complexity FFT-based CD equalization consuming around 20% power in the receiver. In this paper, we propose DFNT-TrDE, an efficient Transform Domain Equalization (TrDE) method that reduces the computational complexity of the CD compensation by leveraging the Fermat Number Transform (FNT, where Fermat number $F_{n} = 2^{b}+1=2^{2^{n}}+1$ ) with diminished-1 representation. We adopt various techniques in the system design. Specifically, we propose High-Radix (HR) FNT to further reduce the complexity for large transform lengths. Moreover, we compare the complexity of 1D circular convolution between 1D-R2, 1D-HR, 2D-R2 and 2D-HR FNT at the granularity of adder level. We furthermore provide recommendations for radix and dimension settings tailored to different transform lengths. The results of our implementation show that the DFNT-TrDE (with $b=8$ ) achieves 62% complexity savings compared to 12-bit split-radix FFT-FDE at a similar Bit Error Rate (BER). The DFNT-TrDE (with $b=16$ ) also achieves 51% complexity savings compared to 16-bit split-radix FFT-FDE at a better BER.
1 December 2025 in IEEE Transactions on Signal Processing
Convolution is fundamental in digital signal processing across many applications. Existing works enable N-point linear convolution via N-point right-angle circular convolution (RCC) based on weighted transforms, effectively removing the need for zero padding. However, these methods are constrained by their choice of weights, which impacts the complexity of weight-related multiplications at the start of the transform and the end of the inverse transform. This limitation leads to either reduced accuracy or increased complexity in weighted Fourier transform (WFT)-based convolution, as well as restricted transform lengths in weighted Fermat number transform (WFNT)-based convolution. In this work, we address these challenges by merging the multiplications by weights into the butterfly structures with arbitrary power-of-2 radix for both WFT and WFNT. We also propose an extraction method to accommodate negative and complex numbers. Our work ensures that weights do not increase complexity, thereby improving accuracy and reducing complexity in WFT-based convolution, while allowing for a broader range of transform lengths in WFNT-based convolution.
14 April 2026 in Optics Letters
ASAbdurrashid Ibrahim Sanka
Abdurrashid Ibrahim Sanka
+1
Chromatic dispersion (CD) compensation is a critical, yet power-intensive operation in coherent optical communication systems. Transform-based methods, such as the FFT and FNT, often rely on overlap-save (OS) or zero-padding overlap-add (ZP-OA) techniques for blockwise equalization in the receiver’s DSP. However, these methods only achieve ( N − N Overlap ) effective points with N -point transforms, limiting the efficiency. In this Letter, we propose the non-zero-padding overlap-add (NZ-OA) method, which achieves N effective points using weighted Fourier transform (WFT) and weighted Fermat number transform (WFNT). The NZ-OA method reduces computational complexity by 41.7% in the 400 km DFT/WFT scenario, and by 15.7% in the 50 km FNT/WFNT scenario when the non-prime Fermat modulus q =2 32 +1 is employed.