Sparse deconvolution using adaptive mixed-Gaussian models
Signal Processing
Sparse LMS for system identification
ICASSP '09 Proceedings of the 2009 IEEE International Conference on Acoustics, Speech and Signal Processing
An affine scaling methodology for best basis selection
IEEE Transactions on Signal Processing
On gradient adaptation with unit-norm constraints
IEEE Transactions on Signal Processing
IEEE Transactions on Information Theory
Hi-index | 0.08 |
In order to improve the sparsity exploitation performance of norm constraint least mean square (LMS) algorithms, a novel adaptive algorithm is proposed by introducing a variable p-norm-like constraint into the cost function of the LMS algorithm, which exerts a zero attraction to the weight updating iterations. The parameter p of the p-norm-like constraint is adjusted iteratively along the negative gradient direction of the cost function. Numerical simulations show that the proposed algorithm has better performance than traditional l"0 and l"1 norm constraint LMS algorithms.