Analytical model of the CKC-based activity index variance
Rok Istenič, Damjan Zazula · International Conference on Mathematical methods, Computational techniques and Intelligent systems · 2008
In this paper a new technique for estimating the number of signal sources is introduced. It is based on the convolution kernel compensation method, in particular on the variance of the activity index. An exact analytical model was build to describe the relationships between the activity index variance and the parameters of the modelled process, such as the number of sources, duration of source responses, and noise corruption. Simulation studies have shown that the activity index variance can be used for this estimation, while it changes with respect to the number of active sources in the signal in a known way incorporated in the derived analytical model. The obtained results are presented and discussed.