Recursive constrained generalized maximum correntropy algorithms for adaptive filtering
- Publisher:
- ELSEVIER
- Publication Type:
- Journal Article
- Citation:
- Signal Processing, 2022, 199
- Issue Date:
- 2022-10-01
Embargoed
Filename | Description | Size | |||
---|---|---|---|---|---|
Recursive constrained generalized maximum correntropy algorithms for adaptive filtering.pdf | Accepted version | 5.91 MB |
Copyright Clearance Process
- Recently Added
- In Progress
- Open Access
This item is currently unavailable due to the publisher's embargo.
Thanks to the ability of preventing the accumulation of errors, constrained adaptive filtering (CAF) algorithms have been widely applied. However, in practice, non-Gaussian noise may significantly degrade the filtering performance of CAFs derived from the second-order signal statistics. In this paper, we propose several constrained generalized maximum correntropy (CGMC) algorithms to overcome this problem, inspired by the robustness and flexibility of GMC to non-Gaussian noises. We first introduce a CGMC algorithm based on the gradient method. To improve its convergence rate with correlated inputs, we further propose a recursive CGMC (RCGMC) algorithm. For RCGMC, we conduct the convergence analysis, and characterize the theoretical transient mean square deviation (MSD) performance. Furthermore, we derive a low-complexity version of RCGMC by using the weighting method and the leading dichotomous coordinate descent (DCD) algorithm. Simulation results demonstrate the effectiveness of our proposed algorithms in non-Gaussian noise environment, and the consistency between the analytical and simulation results.
Please use this identifier to cite or link to this item: