Fundamentals of statistical signal processing: estimation theory
Fundamentals of statistical signal processing: estimation theory
Natural gradient works efficiently in learning
Neural Computation
High-order contrasts for independent component analysis
Neural Computation
Analysis of sparse representation and blind source separation
Neural Computation
A maximum likelihood approach to single-channel source separation
The Journal of Machine Learning Research
Learning Overcomplete Representations
Neural Computation
Underdetermined blind source separation based on sparse representation
IEEE Transactions on Signal Processing
Fast and robust fixed-point algorithms for independent component analysis
IEEE Transactions on Neural Networks
Hi-index | 0.09 |
Unlike the traditional overdetermined or determined blind source separation (BSS), mixing matrix estimation and source recovery are not always coincident in underdetermined BSS. Aimed at the problem of single channel BSS, which is the extreme situation and most difficult case in underdetermined BSS, this paper proposes a mixing vector estimation algorithm for angle modulated signal sources based on a cumulant system of equations. The principle is that the mixing vector can be estimated by solving a cumulant system of equations derived from the probability density functions of the sources. This algorithm does not require specific information about the sources' modulation types or parameters and can be theoretically applied in conditions with any number of sources, even if their frequencies overlap. Experimental results show the estimation performance.