Abstract
A previously proposed model for non-stationary signals is
extended in this contribution. The model consists of mul-
tiple time-translated locally stationary processes. The opti-
mal Ambiguity kernel for the process in mean-square-error
sense is computed analytically and is used to estimate the
time-frequency distribution. The performance of the kernel
is compared with other commonly used kernels. Finally the
model is applied to electrical signals from the brain (EEG)
measured during a concentration task.
extended in this contribution. The model consists of mul-
tiple time-translated locally stationary processes. The opti-
mal Ambiguity kernel for the process in mean-square-error
sense is computed analytically and is used to estimate the
time-frequency distribution. The performance of the kernel
is compared with other commonly used kernels. Finally the
model is applied to electrical signals from the brain (EEG)
measured during a concentration task.
| Original language | English |
|---|---|
| Title of host publication | [Host publication title missing] |
| Publisher | IEEE - Institute of Electrical and Electronics Engineers Inc. |
| Publication status | Published - 2013 |
| Event | 21st European Signal Processing Conference (EUSIPCO 2013) - Marrakech, Marocko, Marrakech, Morocco Duration: 2013 Sept 9 → 2013 Sept 13 |
Conference
| Conference | 21st European Signal Processing Conference (EUSIPCO 2013) |
|---|---|
| Country/Territory | Morocco |
| City | Marrakech |
| Period | 2013/09/09 → 2013/09/13 |
Subject classification (UKÄ)
- Probability Theory and Statistics
Free keywords
- Time frequency analysis
- Locally stationary process
- Optimal Ambiguity kernel
- EEG.