Information fusion via signal fission is addressed in the framework of empirical mode decomposition (EMD). In this way, a general nonlinear and non-stationary signal is first decomposed into its oscillatory components (fission); the components of interest are then combined in an ad hoc or automated fashion to provide greater knowledge about a process in hand (fusion). The extension to the field of complex numbers C is particularly important for the analysis of phase-dependent processes, such as those coming from sensor arrays. This allows us to combine the data driven nature of EMD with the power of complex algebra to model amplitude-phase relationships within multichannel data. The analysis shows that the extensions of EMD to C are not straightforward and that they critically depend on the criterion for finding local extrema within a complex signal. For rigour, convergence of EMD is addressed within the framework of fixed point theory. Simulation examples on information fusion for brain computer interface (BCI) support the analysis. © 2008 Springer US.
CITATION STYLE
Mandic, D., Souretis, G., Leong, W. Y., Looney, D., Van Hulle, M. M., & Tanaka, T. (2008). Complex empirical mode decomposition for multichannel information fusion. In Signal Processing Techniques for Knowledge Extraction and Information Fusion (pp. 243–260). Springer US. https://doi.org/10.1007/978-0-387-74367-7_13
Mendeley helps you to discover research relevant for your work.