.png){kind=icon}
.png){kind=icon}
Return
Diffusion complex-valued least mean kurtosis adaptive algorithm for distributed networks
.png){kind=icon}
PRE
AI
DOI：10.1016/j.jfranklin.2026.108587.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
In the era of big data and increasing system complexity, distributed algorithms have become increasingly significant across diverse fields. Notably, there is a shortage of algorithms for processing complex-valued signals in the diffusion domain. This paper fills this gap by proposing a diffusion complex-valued least mean kurtosis (D-CLMK) algorithm, which uses the negative kurtosis of the error signal as the cost function at each node in the distributed network. This approach offers significant advantages in handling sub-Gaussian noise. Using Isserlis’ theorem, we conduct theoretical analyses of the proposed D-CLMK. Simulation results demonstrate that when applied to complex-valued system identification in distributed networks, the D-CLMK algorithm exhibits a lower mean square deviation (MSD) in the steady state compared with related algorithms. The adoption of diffusion strategies enables the estimation process to achieve higher accuracy, which is critical for practical applications in fields such as sensor networks and Internet of Things (IoT) systems.
Keywords:
Diffusion algorithms
Complex-valued signals
Least mean kurtosis
Distributed networks
Adaptive algorithms
Journal
J
IF:
4.2
Papers:
822
Citations:
0
Organization
No organization information available