.png){kind=icon}
Return
Low Complexity and Robust Diffusion Affine Projection Algorithms for Distributed Estimation
.png){kind=icon}
PRE
AI
DOI：10.1109/TCSII.2021.3127464.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Diffusion affine projection algorithms have the ability to de-correlate the input signal and have faster convergence but with the expense of increased computational complexity. Moreover, traditional diffusion affine projection algorithms consider the noise to be of Gaussian nature. However, practically this noise can be non-Gaussian which can significantly deteriorate the convergence of the algorithms. To mitigate this issue in this brief, we propose two robust affine projection algorithms based on the generalized maximum correntropy criterion (d-A-GMCC) and the logarithmic hyperbolic cosine cost function (d-A-lncosh). To reduce the computational expense of the proposed algorithms, we propose dichotomous coordinate descent based d-A-GMCC and d-A-lncosh algorithms. Extensive simulation study for different Gaussian and non-Gaussian noise environments shows the improved estimation ability of proposed algorithms.
Keywords:
Projection algorithms
Computational complexity
Costs
Cost function
Computational modeling
Kernel
Estimation
Distributed networks
lncosh cost
GMCC criterion
low complexity
DCD
correntropy
affine projection
Journal
I
IF:
4.9
Papers:
8.8K
Citations:
2.5W
Organization
Cited Papers
ShareSaveShareSaveWeighted diffusion continuous mixed p-norm algorithm for distributed estimation in non-uniform noise environment.png)
SIGNAL PROCESSING
SIGNAL PROCESSINGIF3.6
PREAI
ShareSaveShareSave
View more