Bayesian signal detection and source separation in simulated brain computer interface systems
Mubeen, Muhammad Asim · 2016
The problems of signal detection and source separation are important in many fields of science and engineering. In many cases, a target signal needs to be detected in real time and is contaminated by noise. Sometimes the level of noise is on the order of the signal itself. The real time detection of a target signal is of key importance in problems such as the brain computer interface systems. In brain computer interface systems, the neural activity (electric signals) of the brain is detected using sensors (electrodes) on the surface of the brain or the scalp. This signal is contaminated by various types of noise. The level of contamination increases when signal is recorded non-invasively. To detect such signals of interest a Bayesian signal detection technique has been developed and tested for various noise levels and compared with the popular technique of cross-correlation. Receiver operator curves (ROC) are employed to test the robustness of the proposed method and for comparison purposes.