.png){kind=icon}
Return
Efficient Multidimensional Parameter Estimation Using Machine Learning-Assisted SAGE Algorithm
.png){kind=icon}
PRE
AI
DOI：10.1109/LSP.2025.3588079.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Multidimensional parameter estimation is a critical challenge in fields such as radar, sonar, and wireless communications, where the space-alternating generalized expectation-maximization (SAGE) algorithm is commonly used to estimate multipath channel parameters. This paper proposes a machine learning-assisted SAGE framework that integrates Gaussian process regression (GPR) with a two-stage optimization strategy to enhance estimation accuracy and reduce computational complexity. Using a GPR surrogate model, the proposed algorithm streamlines the computation of the likelihood function, significantly lowering processing demands. Additionally, a coarse-to-fine search mechanism mitigates grid mismatch issues. The simulation results demonstrate that the algorithm efficiently estimates the six-dimensional multipath channel parameters, delay, two-dimensional angle of departure (2D-AoD), two-dimensional angle of arrival (2D-AoA), and amplitude, outperforming traditional methods in both accuracy and computational efficiency.
Keywords:
Machine learning
SAGE
optimization
parameter estimation
Journal
IF:
9.6
Papers:
1.1W
Citations:
1.7W
Organization
Cited Papers
ShareSaveChannel parameter estimation for millimeter-wave cellular systems with hybrid beamforming.png)
SIGNAL PROCESSING
OAAI
SIGNAL PROCESSINGIF3.6
OAAIShareSaveShareSaveShareSave
ShareSave
View more